| | r||~, |1||‘ i o _,_;nl.-.-'!:r;' | CENTRAL RESEAR CH L

gy il co RN oL LT
C-84 - Reactors - Special Features - 1 =

3 445k 03b1ckT b of Aircraft Reactors Q?' A A E‘i s

AEC RESEARCH AND DEVELOPMENT REPORT i A 5
rm; W

A ENRGIFIT o
BECLASSHLD | B
. AR el 1RES i

CLASSIFICATION CHANGED Taoz 5 Soamv e i _
- = f
ORITY ‘u'r".---Q,-;—.g.-r.._.-.‘.....:fl_:...(fl_é__-_-___

By AUt y

By: L .:fi_xflcfl-_s.;j..__)e:‘_%.:_‘\:fl__ et
A ZERO POWER REFLECTOR-MODERATED @
REACTOR EXPERIMENT AT ELEVATED TEMPERATURE
D. Scott
G. W. Alwang
E. F. Demski
W. J. Fader

E. V. Sandin
R. E. Malenfant

TDT AT
> A K
!_I:._\.'..I—‘J.'
&

 

i

ST R e

"
i
T

oF
i
e

  

 

 

CENTRAL RESEARCH LISK
DOCUMENT COLLECTION

i | - LIBRARY LOAN COPY
% : . DO NOT TRANSFER TO ANOTHER PERSON

If you wish someone else to see this
document, send in name with document
and the library will arrange a loan.

 

OAK RIDGE NATIONAL LABORATORY
operated by

UNION CARBIDE CORPORATION
for the

U.S. ATOMIC ENERGY COMMISSION

s
o
v

FETRICTED T

 
ORNL-2536
C-84-Reactors -Special Features
of Aircraft Reactors

This document consists of 100 pages

Copy ;‘- of 214 copies. Series A

Contract No. W-T405-eng-26

Applied Nuclegr Physics Division

A ZERO POWER REFLECTOR-MODERATED REACTOR EXPERIMENT
AT ELEVATED TEMPERATURE

Dunlap Scott, G. W. Alwang,  E. F. Demski,*

W. J. Fader,* E. V. Sandin,¥* R, E, Malenfant

Date Issued

AUG 11958

 

* Pratt and Whitney Aircraft

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

TTSSSsAnaiannE.

& TNy

3 4456 03klaLg g

 

 

 

 

 

 

 

 
_ i

SUMMARY

An experiment using a full-scale mockup of the core and reflector of
Pratt and Whitney Aircraft Reactor No. 1 (PWAR~l), a reflector-moderated
reactor, was performed at zero power and elevated temperatures in the ORNL
Critical Experiments Facility. The mockup consisted of a 8.l-in.-dia
cylindrical beryllium central region surrounded by a fuel annulus contained
between twe Hastelloy X core shells. The inner shell was 8.5 in. in diameter
and 0.125 in. thick. The ocuter one varied from 21.4k in. in diameter and
0.156 in. in thickness at the midplane to 14.8 in. in diameter and 0.25 in.
in thickness at the ends. This core shell assembly was covered by a 13-in.-
thick beryllium reflector. Coolant passages, but not the coolant, were
mocked up in both beryllium regions. The control and safety rod consisted
of & 2.4-in.-0D x 2.00-in.-ID cermet anmulus ccmposed of T0% nickel and 30%
rare earth oxides contained in a 0.035-in,-thick Inconel jacket which was
guided along the reactor exis by a 2.875-in.-0D Inconel tube with 0.120-in.-
thick walls. The fuel was a mixture of the fused fluoride salts of sodium,
zirconium, and uranium. The critical concentration of the clean reactor was
10.22 wt% U235 at 1258°F. The calibration of the control rod over the range
from the midplane to the tog of the beryllium resulted in an increase in fuel
concentration to 12.2 wt% U230 and an indicated value of $2.5. The specific
mass reactivity coefficient (Amm)/( Ak/k) was 5.90 * 0.5 and was constant
within experimental error over the range of the experiment. The temperature
reactivity coefficient for this mockup, measured between 1200 and 1350°F, was
-0.47¢/CF. The effect of a B1O neutron filter between the fuel annulus and
the beryllium in the entrance duct region was evaluated for the fast leakage,
fission rate distribution, and reactivity. The power distribution through
the fuel region was measured by counting the fission fragment gemms, -ray
activity in small uranium disks positioned in the fuel amnnulus throughout
the experiment. A measure of the neutron flux distribution in the reflector
was made with gold foils. In addition to a presentation of the measurements
this report also includes a discussion of the important engineering design
features and construction problems.

- iji

 

 

 

 
 

PREFACE AND ACKNOWLEDGEMENTS

The elevated temperature experiment for the Pratt and Whitney Aircraft
Corporation Reactor No. 1 (PWAR-l) was performed as a cooperative effort of
PWAC and Osk Rdige National Laboratory. The component design and fabrication,
which were patterned after a highetemperature experiment that had previously
been performed at ORNL, were largely completed by PWAC at the Hartford,
Connecticut plant under the direction of E. V. Sandin and G. W. Alwang of PWAC
and W. C. Tunnell of ORNL. The final component assembly was carried out by
ORNL personnel under the direction of W. C. Tunnell, and the experiment was
performed in the ORNL Critical Experiments Facility by both PWAC and ORNL
personnel. ‘

Once the reactor was brought to the operating temperature it was kept
at that temperature until all experimental work was completed. This necessi-
tated 24-hr operation, for which personnel was assigned in three shifts as
follows: )

 

 

Chief Reactor Operation Date. Processing Instrumentation
D. Scott E. Demski* M. K, Albright 'V. G. Harness
E. Sandin¥* D. Harvey* R. Malenfant J. F. Ellis
W. Fader* J., J. Lynn G. W. Alwang¥* E. R. Rohrer

In addition, other personnel was available, continuously during the
approach to critical and on short notice thereafter, for special services.
These included J. E. Eorgan, George Nessle, and Jack Truitt of the ORNL
Materials Chemistry Division; R. D. Parten, L. C. Johnson, and W. D. Carden of
the ORNL Health Phiysics Division; and W. F. Vaughn of the ORNL Analytical
Chemistry Division.

* PWAC

iv

 
App. A

II.

TABLE OF CONTENTS

SUmmary « . .+ .+ o« e e . e s . e
Préface and Acknowledgements . . . . .
Introduetion . . . . . . .. .+ . .
Description of Reactor Assembly . . . .

Experimental Results e & s a4 s+ e

-Critical Uranium Concentration . . . .

Control Rod Evaluation . . . . . . .
Effect of B1O Sleeves on the Reactivity .
Critical Concentration as a Function of Rod
and B0 Sleeve Position
The Specific Mass Reactivity Coefficient .
The Effect of Temperature on Reactivity .
Fuel Importance in End Duct . . . . .
Fast-Neutron Leakage Measurements . . .
Longitudinal Fission Rate in End Duct . .
Radial Fission Rate Distribution . e e

Gold Neutron Flux Distribution in the Beryllium

Reflector e e s e e e e e e
Engineering Report . . . . . . . .
Description of Assembly Components . . .

Reactor Tank e e e s e e s e
Sump Tank el e e e s e e e e e
Enricher . .« « + =« o & o« « & o
Control Rod Drive . . . . . . .
Source Drive e e o e . e
Gas System and Nuclear Controls . .
Filling and mixing circuit .
Standby heljum supply circuit .
Moderstor helium supply circuit . .
Fuel level indicators .. . . .

Position

Connéctions between gas system and nuclear

controls « .+ .. .+ ¢ ¢ & . .
Nuclear instruments and controls .
Snow Traps . .« +« + 4+ o * s+ e
Heating and Temperature Control Circuits

Construction of Assembly Components s o

Welding . . e e « e« s W
Fgbrication of Core Shells e o s o e

v

   

Page No.
iii

iv

 

 

 
 

Fabrication of Blo

Fabrication of Gold Foils

Sleeves .
Fabrication of ByC-Cu Plates
Fabrication of Beryllium Parts

Fabrication of Control Rod .

Fuel Manufacture o .
Material Failures . .

Typical Sequence of Operation

Enriching . . .« .« .
Mixing e« + e« s+ e
Sampling . « » « s
Instrument Checks . .
Filling the Reactor Core
Temperature Readings .

Approach to Critical (Routlne)
Maintaining the Power level

Shutdowvn .« .« =« « o

App. B. Composition and Weights of Reactor Materials

App. C. List of Figures . e e

App. D. List of Tables . . . .

 

 

&
O
a

INTRODUCTION

- A reflector-moderated reactor experiment was performed at elevated
temperatures at the Critical Experiments Facility of the Oak Ridge National
Leboratory (ORNL) as part of the circulating-fuel reactor program of the
Pratt and Whitney Aircraft Campany (PWAC). In general, the purpose of the
experiment was to experimentally verify the theoretically predicted nuclear
properties of a PWAC reactor (designated as PWAR-1l) and, in addition, to
establish those properties of the reactor system which are not susceptible to
accurate theoretical analyses. Prior to the experiment some effort had been
expended in correlating PWAR-1l analytical work t¢ the room-temperature and
elevated-temperature critical experiments performed at ORNL in conjunction
with the Aircraft Reactor Test (ART).l Owing to the similarity of the design
between the ART and PWAR-1l, this was of considerable value in esteblishing
analytical techniques suitable to the PWAR-1l; however, it was recognized that
before absolute confidence could be placed in these techniques the results
should be verified by an experiment designed especially to mock up the PWAR-1.
The choice of an elevated-temperature experiment, as opposed to a more
versatile room-temperature experiment, was largely based on the following
considerations:

1. The over-all PWAR-1 design was considered to be well established,
since it was possible to check analytical procedures against the
series of ART roam~temperature experiments.

2. The information still to be determined, therefore, could be
obtained better in an elevated temperature mockup which not
only conformed more exactly to the design in temperature,

- but also in geometry and in the physical composition and
properties of its component materials.

5+ The ease with which uranium concentration could be increased
in an elevated-~temperature experiment made it more feasible
from an operational point of view.

~ The decision to perform the elevated-temperature experiment was made
early in January, 1956, after discussion between representatives of both
PWAC and ORNL. At that time the responsibility for design and fabrication
of most of the components was assigned to PWAC. Accordingly, advanced
estimates of material requirements were made and by the end of February, 1956,
essentially all materials~had been. ordered. At the same time preliminary
drawings of the moderator had been supplied tp the Brush Beryllium Company
and fabrication of the moderator was begun. Engineering design work was
continued and final layouts were campleted late in March, 1956.

By May 8, 1956, all detail drawings had been released. It was hoped that
< the fabrication and assembly of all ‘components could be campleted by July 1,
1956, but, because of exceptionally long delivery times for nickel-based alloys

 

- 1. The ART has been described in various progress reports, see, for example,
"ANP Quar. Prog. Rep. for Period Ending Sept. 10, 1955," ORNL=1947, p. 58;
"for Period Ending Dec. 10, 1955," ORNL-2012, p. T1.

Gy

 

 
and many problems encountered in fabricating the core ghells and the
moderator, the component assemblies of the experiment did not arrive at ORNL
until November 9, 1956. -

Final setting up of the experiment began immediately. A variety of
problems were encountered during this period and the experiment was not
ready for the introduction of fuel and the beginning of multiplication
messurements until Jamuary 28, 1957. A sumary and description of the
problems which occurred during fabrication and final assembly are contained
in Appendix A. The experiment was made critical on February 5, 1957, and by
the end of February all data had been taken and disassembly had begun.

The experiment was run at essentially zero nuclear power. The operating
temperature was held constant at approximately 1250°F, which corresponds
closely to the design operating temperature of the PWAR-1 moderator; this
temperature was maintained by external heaters.

Dufing the course of the experiment various nuclear parameters were
investigated as requested by PWAC.

] . T
oW oy
3y L.

7 3'1[*"‘"

 

 
I. DESCRIPTION (F REACTOR ASSEMBLY

The apparatus consisted of the following major components (see Fig. 6):

the
.the
the
the

reactor assembly,

reactor tank,

sump tank,

enricher,

the control rod drive, source drive and supports,
the inert gas system, " |
nuclear control and instrumentation circuits,
heating and temperature control circuits.

- . -

A .

O~ O\ £\ N

The reactor assembly was mounted within the reactor tank in a helium atmos-
phere. The sump tank, located under the reactor tank, contained the fuel when
it was not in the reactor core. BHelium pressure was used to transfer the fuel
from the sump tank up through a fill and drain line to the reactor core. The
fuel consisted of the ternary system NaF-ZrF)-UF) which has a liquidus temper-
ature in the neighborhood of 1000°F. Additional uranium in the form of NapsUFg
was added to the fuel in the sump tank by means of the enricher.

After criticality was achieved a chemical analysis was determined after
every third enrichment. A continuous inventory of the contents of the sump was
maintained and was used to indicate the uranium concentration when a chemical
analysis was not available. Table 1 gives the weight percent of each component

 

as obtained by the two methods and demonstrates that the calibration of the
Sample 12 was the initial critical concentration.

enricher. wag reasonable.

 

 

 

Table 1. Fuel Constituents
Uranium (wt%) UFy (wtb) NeF (wth) ZrFy (wt%)
By By By By By _ By By By
Number iAnalysis| Inventory|Analysis|Inventory|Analysis [Inventory| Analysis |Inventory
12 10.97 10.963 | 14.514 | 1k.501 20.082 20.152 | 65.416 | 65.343
13 11.13 11.066 14 .640 20.154 65.206
14 11.26 11.180 14 .790 - 20.155 65.054
15 11.39 11.310 14.969 20.157 6L4.873
16 11.52 11.450 15.148 20.159 64 .692
17 11.64 11.58% 15.325 20.161 64.513
18 11.8% 11.718 15.502 20.163 64.335
19 11.96 11.851 15.678 20.165 64.156
20 12.09 11.983 15.853 20.166 63.979
21 12.20 12.137 16 .057 20.169 . 63.773
22 12.52 12.312 | 16.563 | 16.288 20.453 20.171 | 63.050 | 63.540

 

 

 

 

 

 

 

 

 

 

 
 

The configuration of the reactor assembly is shown in Fig. 1. The fuel
was contained in the annulus formed by the inner and outer Hastelloy X core
shells. Incremental volumes in this core are given in Table 2. The outer
shell was 0.156 in. thick in the region of the midplane (cross section at
meximim core Adiameter) and 0.250 in. thick in regions further than T- 3/4 in.
above and below midplane. The inner core shell was uniformly 0.125 in, thick
except in the lower end duct where its thickness increased to 0.250 in. Due
to difficulties encountered in the fabrication of these shells, variations
in the wall thickness of #0.025 in., as well as out-of-roundness and de-
partures fram designed contour, occurred. Appendix A gives an account of
these difficulties plus a detailed picture of the final shapes and thicknesses.

The central moderator column, or the island, consisted of beryllium through
which longitudinal holes 0.250 in. in diameter were drilled to mock up coolant
passages. The distribution of these holes is given in Table 3. An annular void
0,125 in. in average thickness was located between the inner core shell and the
island beryllium to mock up a coolant passage. Because of engineering and
operational ccomplications, none of the coolant passages contained sodium as they
would in the power reactor itself. A 2.935-in.-dia longitudinal hole through -
the center of the island contained. the control rod thimble, a 2 B875=in.-0D
Inconel tube with 0.120-in.-thick walls. A 12.5-in.-long beryllium cylinder,
contained in a stainless steel jacket, was placed in the bottom of the thimble
to limit the lower travel of the rod in the event of mechanical failure in the
rod support linkage. |

The control rod, which also served as a safety rod, con31sted of a
2.430-in.-0D x 2. 000-in.-ID anmlus of T0% Ni, 30% Lindsay Mix® cermet 36 in.
long (see Appendix A). This annulus was physically contained in a 0.035-in.-
thick Inconel jacket. Its motion for control purposes was confined to the
region fram 24.155 in. above to 1.415 in. below the midplane.

A k.3 x lO6 neutrons/sec (Jan. 31, 1957) Po-Be source rode longitudinally
down through the c¢enter of the annular control rod into the control rod thimble
to its normal position which, during multiplication measurements and startups,
was a little below the reactor midplane.

The reflector was built up around the core shell assembly, Fig. 2, with
Yoin,-thick beryllium rings as shown in Figs. 3 and 4., Holes 0,250 in. in
diameter were drilled through these slabs to approximate the probable distri-
bution of coolant passages in the reflector of the power reactor. The distri-
bution of these holes at three elevations as a function of radial distance from
the reactor axis is shown in Table 3. A coolant anmulus having an average
thickness of 0.125 in. was also provided between the outer core shell and the
reflector beryllium, Again no sodium was contained within these coolant
pessages; however, & helium atmosphere was maintained in the reactor tank as
shown in Fig. 5 to protect the beryllium. For the radial gold foil acti=-
vation measurements, foil holders were provided which were placed into slots
in the reflector at three elevations. The complete assembly, with its
covering insulation, is shown in Fig. 6.

 

* Lindsay Mix is a rare earth oxide mixture consisting of 63.8 wt% Sm,
26.3 wt% Gd, L. 8 wt% Dy,and 0.9 wth Nd.

 
 

 

 

 

 

 

   

 

 

 

 

 

 

4R
ORNL-LR-DWG 29826
CONTROL ROD THIMBLE
CONTROL ROD
BELLOWS EXPANSION JOINT
I~y
" RE-ENTRANT TUBE—\ " FILL PROBE
M I
’ | T
- - | |l
1&‘
1l |‘
I (ke
OUTER AND INNER j \“
BORON SLEEVES i
!i TOP PLATE REACTOR TANK
al‘_ I
i
N ‘ i[
INNER CORE SHELL BERYLLIUM ISLAND
OUTER CORE |
SHELL 5
D 5
4 BERYLLIUM
REFLECTOR
DETECTOR
TUBES
FUEL A
NUL | .
,, | FNTOLes W 8 _ MID-PLANE
B
b,

BERYLLIUM PLUG

 

b
‘ Oy BOTTOM PLATE REACTOR TANK

 

  

- LOWER B,C-Cu PLATE OUTER AND INNER
8,C-Cu RINGS

FILL AND
DRAIN LINE

 

FIGURE 1. REACTOR ASSEMBLY -~ MAJOR COMPONENTS

 

 
 

 

Teble 2. Incremental Core Volumesa
7.0ne Voll.tlnE;b Zone Volume
(in. above midplane) (in.2) (in. below midplane) (1n.”)
26 to 25 115.14 0tol 292.87
25 " 24 115.19 1 "2 288.24
24y " 23 115.23 2 "3 280.06
23 " 22" 115.33 3 "4 270.69
22" 21 115.47 L "5 260.19
21 " 20 115.70 5 "6 247.70
20" 19 116.40 6-" 7 233.94
19" 18 117.57 7 "8 219.61
18" 17 119.44 8 "9 205.65
17" 16 121.57 9 " 10 191.48
16" 15 124 .66 10 " 11 176.91
15" 14 131.16 11 " 12 162.50
" 13 140,26 12 " 13 149.58
13" 12 150.60 13 " 14 139.27
12" 11 162 .24 s " 15 130,43
11" 10 175.01 15 " 16 124.18
10" 9 189.82 6 " 17 119. 4k
v 8 204,50 17 " 18 119.97
8" T 218.73 18 " 19 115.02
"6 233.33 19 " 20 117.00
6" 5 247.39 20 " 21 119.00
"o 259.56 21 " 22 120.80
L "3 269.86 22 " 23 123, bk
3" 2 279408 23 " a2k 155.01
2" 1 286 .60 24 " 25 117. Tk
1" 0 292 .54 25 " 26 85.99
26 " 27 83.71

 

 

a. Fach core volume quoted for a l-in.~high zone was obtained by numerical
integration of the final core dimensions. The total volume to 2 in.,
sbove the beryllium obtained by integration was 9159 in.? or 150.2 liters;
the measured volume whs.14Ta2-liders. - :

b. These volumes correspond to- an unheated assembly. The volumes for an
gssembly at 1250°F are obtained by multiplying by 1.032.

 
Tsble 3. Distribution of Holes® in the Reflector and Moderator

 

 

 

 

 

 

 

| f In Reflector - . - L
16 in., Above and Below 8 in. Above and Below In Island | e
Midplane _ Miiplane " . At Midplane. at All Elevations
Redius, No. of sb © No. of Radiusb No. of Redius, No. of
(m.) Holes - (in.) Holes (in.) Holés (in.) Holes
8.123 120 - 9.897 120 11.134 120 1.787 16
8.626 60 10.400 60 11.637 60 2.137 16
8.710 60 10.484 60 11.721 60 2.687 - 32
9,162 60 10,900 60 12.137 60 3,287 32
9,234 60 -11.067 60 12,304 60 3.687 32
9.793 60 ~11.567 60 12,80k 60 ‘
10,126 - €0 11,900 60 13.137 60
10.626 60 -12,400 60 13.637 60
11.126 60 -12.900 60 14,137 60
11.626, 60 13,400 60 14,637 60
12.293 60 14,067 60 15.304 60
12.960 60 14 o T3Y 60 15.971 60

14,626 60 - 16.400 60 17.637 60

 

 

. All holes were O, 250 in. in diameter and were distributed equa.lly around 560 d.eg at the
given radius.

be. The radii given apply only at the indicated elevations since these are continuous hojles
which roughly follow the ocuter core shell contour. Straight lines between corresponding
hole ra.dii describer the path.

 

 
 

 
  
 
  
    
  

ST
PHOTO 27699

 

 

 

  
  

 

| -
l |
-<+— CORE SHELL ASSEMBLY
H :
= o
|

BOTTOM PLATE ®
REACTOR TANK

   

K

 
     

FILL & DRAIN LINE |
‘J =i el _ ‘ - g

    

S

[t |

  

)

 

 

 

Fig. 2. Core Shell Assembly in Place.

 

 
-
PHOTO 27700

L r

d

A

LS - SR <— BERYLLIUM
COOLANT CHANNELS : o) ' ; g REFLECTOR £
-ty
-t

N

BOTTOM PLATE
REACTOR TANK

t _'_-

.1;':

 
 

q OUTER_BORON
' SLEEVE-—-— <

 

 

A
THERMO COUPLE l

. B l i . .
-‘ ,a\'.*_, 1‘_{ ,.-
\ H. X f. " L : <
i&f X

BOTTOM PLATE |8
REACTOR TANK (i

L

Fig. 4. Reactor Assembly Complete.

 

 

 
UNCLASSIFIED
PHOTO 27830

UPPER SNOW TRAP

[/

 

N
BN~ REACTOR TANK

 

 

LOWER SNOW TRAP

 

   

 

 

 

Fig. 5. Reactor Tank in Place with Heaters.

 
 

_12_

, ' UNCLASSIFIE[q
PHOTO 28232 I

]"

 
  
 
  
     
  

{'<— CONTROL ROD AND
_®&\ - SOURCE DRIVE

..-.-fll “
i 0

.‘.

 

  
  

 

@

    

—_— .

 

 

ENRICH ER \\w

P Sie i

Fig. 6. Assembly Complete.

 

 
15

Ratural boron carbide-copper plates, which.simulated'boron-containing
regions in the PWAR-1 design, surrounded the lower end duct. A bordh carbide-
copper plate was also placed at the bottom of the island to limit fissioning
to the core region. The areal density of boron carbide in these pldtes was

0.35 g/cmC.

1 Sleeves containing elemental B1O powder packed to a density of 1.5 S/Cc
in an annulus 0.060 in. thick surrounded the upper end duct. These sleeves,
which acted as neutron shields, could be moved from a position of complete
retraction to their fully inserted position about 6 in. below the top of

the beryllium by actueting rods protruding from the top of the reactor
through gastight fittings. Longitudinally through the end duct anmlus, in
the neighborhood of these shields, a re-entrant tube was provided in which
fission rate measurements were made for various positions of the boron
shields. '

It will be noted that while the reflector shapes in the power reactor
consisted of continuous curves, these curves were approximated by straight-
line segments and step functions in the experiment. The sketch shown in
Fig. 7 gives a comparison between the power reactor and the critical experi-
ment. o ’

The level of the liquid in the core was normally held 1iin. above the
top of the beryllium moderator. The safety system was arranged to dump the
fuel into the sump tank under various hazardous conditions, muclear or
mechanical, which will be described later. It required 1.8 to 2.0 sec for
the fuel to drop the first 6 in. and 29 to 30 sec to campletely empty the

Detailed dimensions of the reaetor assembly relevant to interpretation
of the experimental results are given in Fig. 8 and Table 4. Appendix A
contains, in addition to a discussion of all other components of the experi-
ment, a further account of the engineering problems associated with the"
reactor assembly. Appendix B gives analyseg of the materials contained in
the assembly as well as weights of the moderator and core shell components.

11. EXPERTMENTAL RESULTS

Introduction

Upon the request of PWAC, the following experimentel informstion was
dbtained:

l. Criticel concentration as a function of control rod position.

&. Reactor period for small displacements of the control rod.

5. The effect on control rod worth of increasing the areal density of

- rod poison. .

4, Critical concentration as a function of the position of BLC
shields around the upper end duct.

 

 

 
 

 

G
ORNL-LR-DWG 29827

vzl CORE SHELLS | pmomes=eomeee- -
REACTOR BERYLLIUM

REACTOR BORON

------- CRITICAL EXPERIMENT BERYLLIUM

CERIEENE CRITICAL EXPERIMENT BORON q--—-

   
     

  
 
 
  

o e A et s B M T M e D S m S e e e e
- ———— -

- i e mm e SE W MR R W G b S M WA W S e

e e m m e m el A

—= \
== \\\\\\\\
'N

 

-
—

TR R TR PP
e fudln® ok el o'’ " S

  

| v e e e s mw e o e o e o mm ey e yw W s e e sm S iSO S

 

i
b3

FIGURE 7. A COMPARSION OF THE POWER REACTOR DESIGN (PWAR-i) AND THE ELEVATED TEMPERATURE CRITICAL ASSEMBLY
5. The over-all temperature coefficient of reactivity between isothermal
states in the region between 1200 +to 1350°F.

6. Leakage of fast neutrons fram the upper end duct for various positions

~of the B1O ghields. | |

T. Longitudinal power_distribution in the upper end duct for various
positions of the B1O ghields.

8. Radial power distribution at three elevations in the core and at one
position in the lower end duct. : |

9. Radial neutron flux distribution at three elévations in the moderator
as indicated by gold foil activation.

Critical Uranium Concentration

The clean critical uranium concentration is defined as that concentration
of uranium in weight percent of a specified fuel for which, at a specified
temperature, the experiment was critical with the control rod and BlO gleeves
campletely withdrawn. This concentration was approached stepwise by additions
of fluoride salt mixtures of high UF} concentrations to the mixtures in the
fuel sump.

During the approach to criticality both B0 sleeves were fixed in their
full-out positions, and the temperatures of the fuel sump and reactor were
approximately 1250°F. The initial content of the sump was 523.7 kg of a
N&F-Zth-UFh mixture, with g uranium concentration of h.35 wt%. Previous
experience with ART mockup critical assemblies at room temperature and at
12009 indicated that this concentration would be far below the critical
concentration. The fuel was raised into the reactor core to the level of the
fill probe, the control rod was withdrawn, and the neutron counting rates of
two BFz long counters and one Hornyak sgintillation counter were determined
while The Po-Be neutron socurce (4.3 x 10 neutrons/ sec) was at the reactor
midplane.

About 13.8 kg of NapUFg containing 59.8 wt% uranium was then added to the
fuel sump. The fuel mixture was homogenized by raising it to the level of the
reactor midplane and dropping it back into the sump 15 times. To check the
effectiveness of this mixing procedure, a sample of the fuel was extracted for
chemical analysis after 10 mixings and asgain ‘after five more mixings. The
reported analyses of the uranium concentrations of the two samples differed by
1% of the value, which was 5.86 wt% uranium. The probable error reported was
+0.2% of the value. The fuel was then raised to the level of the £i11 probe,
the control rod was withdrawn, and the counting rates of the neutron counters
were determined while the Po-Be source was in the reactor as before. Four
more batches of 12.6 to 13.8 kg of Ne,UFg were added to the fuel sump. After
each addition the fuel was mixed 15 times, a sample was extracted for
chemical analysls, and the neutron counting rates were determined as above.
The four additions resulted in a uranium concentration of 10.Th4 wt% by chemical

analysis,

 

 

 

 
 

16

Table 4. Dimensions for Réactor Assembly Sketch

 

, Cold Hot .
Designation® (in.) (in.) Description
A 23%.906 2} ,155 Contrpl rod ‘zero
B 25.855 2h.155 Midplane to top of island beryllium
C 24,250 2L4.553 Midplane to top of reflector—béryllium
D - é5.527 Midplane to top of fuel, fill probe on
E 17.757 17.924 Midplane to lover limit of longitudinal power
' traverse :
F 25,649 25.927 Mid.plané to top of inner B)C-Cu ring inside lower
end duct ‘
G 23,750 24.028 Midplane to top of outer BhC-Cfi ring around lower
end duct N
H 14%.812 14.949 Midplene to top of heryllium plug in lowefi portion
of control rod thimble |
T 26.750 27.028 Midplane to base of reflector beryllium :
K 27.149 27.427 Midplane to base of island beryllium |
P 10.529 10.639 Meximum inside radius of the ofiter core shell ‘
M 24,000 24.274  Redius of reflecto£ (sidb G)
N 4.000 4.0M6 Radius of island beryllium
P 1,318 1.332 Radius of control rod thimble
R 16,000 16,182 Radius of reflector (slsb B)
S 8.920 9.013 Redius of outer core shell in lower end duct.
T 7.250 7.325 Radius of imner core shell in lower end duct

 

a. Refer to Fig. 8.
b. Arrow displaced for clarity.

 
 

<
ORNL-LR-DWG 29828

 

 

 

_MID-PLANE

 

 

FIGURE 8 REACTOR ASSEMBLY ~-DIMENSIONS

 
 

18

Subgequent additions of uranium were made from the enricher in increments
of about 750 g of NapUFg containing 59.8 wt$ uranium. The procedures of
mixing, sampling and neutron counting rate determinations were carried out
after each addition. In Fig. 9 the ratios of the neutron counting rates
obgerved st the 4.35 wt% concentration to the counting rates observed at
higher concentrations are plotted versus the uranium concentrations in weight
percent reported from chemical analysis. The reactor first showed a positive
period with the control rod and source withdrawn at a concentration of 10.37 wtb
+ 0.02 wt$ uranium (10.22 wt$ U235). The reactivity corresponding to this period
was 4.63 cents as calculated from the inhour formula with B = 0.0073. The mean
reactor temperature at this point was 1248CF. Using the value of =0,473 cents/OF
for the temperature coefficient of reactivity (p. 32), it was determined that
with the rod withdrawn this concentration would constitute a critical system at
12580F. The probable error is that reported for the result of the chemicsal
analysis. Chemical analyses of the sample containing 10.97 wt$ uranium showed
that other constituents in the mixture were as follows: zirconium, 35.7 t 0.2 wtk;
sodium, 11.0 £ 0.1 wt$; and fluorine, 42.4 £ 0.4 wtb. The corresponding value
of the UF) concentration in mol percent is 5.09 ¥ 0.0l.

Control Rod Evaluation

 

The reactivity worth of the control rod was evaluated over the part of
its travel fram the midplsne to approximately the upper surface of the beryllium
island. The lower end of the rod was used as the reference in indicating the
rod position.

The value of the rod in cents was obtained from period measurements in the
following manner. A small known amount of the enriching salt was added from
the enricher and thoroughly mixed with the fuel. The fuel was then raised into
the reactor and the rod pulled out to its previous position at critical, which
put the reactor on a positive period. At an appropriate power level the rod
was inserted to determine the new position at critical. The period was de-
termined from the strip chart record of the logarithm of the neutron level as
seen by a BFz ionization chamber. The reactivity in cents required to override
this period was calculated using the inhour equation with five delayed neutron
groups and an effective delayed neutron fraction of 0.0073. From these data the
average sensitivity of the rod in cents per inch over this interval of rod
travel was found by dividing the value of the period expressed -in cents by the
distance in inches the rod traveled to override this period. A plot of the
sensitivity as & function of control rod position is given in Fig. 10.

If during control rod calibrations the fuel additions are of an appropri-
ate size so that a reasonsble period, i. e., between 100 and 200 sec, is obtained
when the rod is withdrawn to the pervious critical position, the total value of
the rod can be obtained by summing the values obteined for each successive
enrichment. At least two determinations of the period were made for each en-
richment. Since slightly different critical positions were usually obtained
after each period determination, the rod position used for the next enrichment

 

 
 

 

 

 

 

 

3 .
10 ~ ORNL-LR-DWG 29829
T [ I I I
ADJUSTED BAl
POINT
O COUNTER #
[0 HORNYAK BUTTON
O COUNTER #3
0.8} .
0.6
£ 1
M ¢
04l
0.2t~ -
O
0
00 L 1 i 1 { ] O,
40 50 60 70 80 90 10.0 1.0
URANIUM CONCENTRATION (weight percent)
FIGURE 9. RECIPROCAL MULTIPLICATION

 

 

 
_20_

ORNL-LR-DWG 29830
25

 

TEMPERATURE 1255° F

20~

SENSITIVITY {cents/inch)

MiD-PLANE 24.155"

 

 

 

 

 

 

 

0 / | ] |

0 5 10 i5 20 25 *
DISTANGCE CONTROL ROD INTO BERYLLIUM (inches)

FIGURE 10. CONTROL ROD SENSITIVITY .

 

 

 
-

was chosen either as the average value of the preceding critical positions

- or, in cases where the positions varied a considerable amount, as the last
position obtained. In using these data to obtain the integrated value of the
rod, each period value in cents found for any one enrichment was added to the

- 'wiugiég'the rod for the position used during the period. In most cases the
values so obtained when plotted against rod position defined a curve with a
glope near the value of the sensitivity of the rod determined directly from
the peripd. This Jjustified using a direct interpolation to campute the inte-
grated value of the rod for the position used to put the reactor on a positive
period after the next enrichment, if this did not coincide with a previous
critical position. In some cases the curve through these sums had a slope of
the wrong sign, probably due to temperature changes during the run, in which
case an inverse interpolation was performed to find the integrated value of the
rod for the position used during the succeeding period.

21

In one case the poslition of the rod used to produce & period was outside
the previously calibrated range. For this case, the integrated value of the
rod was found by extrapolation, using the sensitivity determined directly from
the period. This extrapolation was short, 0.2 in. of rod travel.

A plot of the total control rod value in cents as & function of the rod
position is shown in Fig. 11. The data used to plot this curve are listed
in Teble 5. The position of the rod plotted in all cases is the distance the
tip of the rod was inserted below the top of the island beryllium as indicated
on the Veeder Root counter attached to the selsyn repeater on the control panel.
The relation between the Veeder Root indication and the actual position of the
control rod was checked before each run,

The evaluation of the control rod by this method assumes that the tempera-
ture remains constant during the time of the run. The fuel was usually raised
into the reactor region and allowed to remain there about 45 min so that it
would reach thermal equilibrium before the reactor was put on a positive period.
The method of averaging the data to obtain the curve of the integrated value of
the control rod versus position effectively gives a reduced weight to any
measurement for which the error due to changes in temperature during a run is .
large. '

The period produced by setting the rod at or near a previous critical
position is. a result of the net reactivity change which occurred since the
last time the reactor was critical and the value of the period is independent
of what specific changes occurred. Experiments indicated that the sensitivity
of the rod did not change apprecisbly with changes in concentration and
temperature for the range over which these quantities vdried during the course
of the experiment. Therefore, the increment of the rod required to override
the periods produced during rod calibration runs, when only the concentration
and temperature varied, was independent of what changes occurred. For small
changes in temperature the effect of changes in dimensions of the rod due to
thermal expansion was assumed to be negligible. Thus, this method of
. evaluating the rod is independent of temperature changes which occur between

runs. However, in using the curve of total rod value versus control rod
position as a calibration curve to evaluate any specific reactivity change

 
-2 -

CRNL-LR-DWG 29834

 

 

 

 

 

 

25

20

 

 

 

| [ |
GGI'p2 3INVIJ-AIN
w
o
n
n
o
ul
r
2
&
— w
a
=
Ll
T
WNITTAY38 aNvsl 40 dO1
LBBE'0  WNITIAY3E HO103743H 4O 401
Le4El 38084 T4 Ly TIAZIT 13N4 4O LHOIEH
| | _
? 8 3 3
N o - -
D

{Siu=0) 3NTIYA AOH TI0H1INO

DISTANCE CONTROL ROD INTO BERYLLIUM (inches}

CONTROL ROD EVALUATION

FIGURE

 

 
23

Table 5. Reactivity Value of the Control Rod at Various Positions

 

Distance Between lower
End of Rod and Top of

 

Beryllium Rod. Value
(in.) | (cents)
0,000 0.0
5.520 3.80
6.180 6.80
7.960 13.89
8.075 14.10
9.460 21.85
10.660 29.76
11.270 34,95
12,384 45.24
13.136 53.41
14.019 63.86
14.705 72.29
15.420 - 81.9%
16.000 90.14
16.685 100.40
16.809 103.14
17.208 109.89
17.755 118.90
18.285. 127.30
18.726 135.73
19.201 1hh by
19.735 154,84
20,489 169.15
20.970 179.30
21.284 185.16
21.688 193.76
22.090 202.48
22.475 210.90
22,950 221.20
23.350 230.05
24.000 2k ,15

24,155 . (Midplane)

 

which is produced in the reactor, any change in average temperature which
occurs during the time the desired change is made must be, taken into account.
A camplete temperature survey of the reactor, therefore, was recorded during
each run of the experiment, and an average reactor temperature was found by
the method described on page 30.

 

 

 
 

2h

The total effectiveness of the rod extrapolated to the midplane from
the measurements of this experiment was 248¢ or a ak of about 1.8%. This
was considerably less than the value of about 5.8% predicted by multigroup
calculations performed by PWAC. In order to check the effect of increasing
the density of rare earth poisons in the rod as a possible way of increasing
the effectiveness of the rod, one of the control rods used in the ART Hot
Critical Experiment was modified so that it could be inserted inside the rod
used in this experiment. This ART rod had a 30-in.-long annulus of Lindsay
Mix rare earth oxides 1/8-in.-thick with an cutside diameter of 1.275 in.

It was enclosed in an 1.34-in.-0D Inconel tube with 0.020-in.-thick walls.

At the end of the experiment this rod was inserted so that its lower end was
within 1/4 in. of the end of the outer rod. The system was critical when the
combined rod was 3.25 in. from the midplane or 1.7 in. from the previous
critical position. This is about 40¢ or a net increase in effectiveness of
approximately 16%. If this reactivity value is considered as an areal effect
and the rod diameters are used to correct it, the value becomes T6¢, which
gives a 50% increase as an upper limit. While large in magnitude, this is
not enough of an increase to account for the large difference between the
calculated and experimental values.

Effect of B1O Sleeves on the Reactivity

 

The effect on the reactivity of the movable B1O s1ceves was evaluated
for three uranium concentrations and for several degrees of insertion. The
results are listed in Table 6 and a plot of the data is shown in Fig. 12.

The position of the sleeves was determined from notches filed in each of
the three support pins attached to the inner and outer sleeves. These marks
were made during the initial warm-up of the reactor when the sleeves were at
the bottom of the slots provided for them in the island and reflector be-
ryllium. The marks coincided with the top of the collar in the Conax fittings
through which the pins passed. Since the slots in the beryllium were 6 in.
deep, the position of the sleeves listed in Table 6 is the average of the
distagce measured between these marks and the top of the collar subtracted
from in.

Because of difficulty encountered in moving the outer sleeve for the
experiments at the two higher uranium concentrations, two of the three
actuating pins broke loose; it was later discovered during disassembly that
all three pins had been severely bent. Therefore, the position of the outer
sleeve is somewhat uncertain except in the fully inserted position when it
was possible to feel the sleeves strike the bottom of the beryllium slots.
When the sleeves were withdrawn to the "Out" position the distance from the
file marks to the Conax collar was more than 5-1/2 in. in all cases.
Therefore, since it is doubtful that the pins straightened out or stretched
apprecisbly when the sleeves were pulled up, the sleeves in the "Out" position
extended no more than 1/2 in. into the beryllium. For the position "1/2 In"
at a concentration of 12.20 wt% uranium, the file marks were 3 in. from the
collar, but the uncertainty in the position of the outer sleeve is probably
1/2 in. or more.

 
25

Teble 6. Effect on Reactivity of Inserting the BO
Sleeves into the End Duct Beryllium

 

Average Depth

 

of Sleeves in Uranium Loss in
Beryllium Concentration Reactivity
(in.) (wt%) (cents)
0.26 11.52 0 (Reference)
1.3 11.52 2.6
3.0 11.52 16.5
5.5 11.52 64.1
Out 11.83 0 (Reference)
In 11.83 T3.2
Outer Sleeve In, Immer Sleeve Out 11.83 54.8
Inner Sleeve In, Outer Sleeve Out 11.83 27.6
Oout 12.20 0 (Reference)
1/2 In 12.20 12.6
In 12.20 69.6

 

The control rod position for criticality was determined for each position
of the sleeves. An average temperature over the reactor was cobtained from the
thermocouple readings and the rod position data corrected for temperature
changes from run to run. The reactivity sensitivity of the rod at each sleeve
position was determined from period measurements and compared with the sensi-
tivity obtained during the calibration of the rod. The sensitivity had de-
creased slightly, especially for the cases where the sleeves were fully inserted,
but no correction was made. Neglecting this effect overestimates the value of
the sleeves fully inserted by less than 10% for the 11.52 wt% uranium case,
where the effect is greatest, and less than 2% for the 12.20 wt% uranium case.

At a concentration of 11.83 wt% of uranium the effect of each sleeve was
determined by inserting each sleeve separately to the full "In" position
while the other sleeve remained in the "CQut" position and finding the rod
position for critiecality. These results also appear in Table 6 and are
plotted in Fig. 12 and indicate that the outer sleeve is a@bout twice as
effective as the inner sleeve. The interaction of one sleeve with the other
is indicated by the fact that the sum of the values of each sleeve alone is
greater than the effect of both sleeves together by about 15%.

 

 
 

(cents)

CGHANGE IN REACTIVITY

80

_26_
S

ORNL-LR-DWG 29832

 

70—

 

 

| | |

O 1152 WT % U
A 1183 WT %U
O 12.20WT. %U

A OUTER
SLEEVE
IN
INNER ™ |
SLEEVE
ouT

A INNER
SLEEVE
IN
OUTER

SLEEVE__
ouT

 

 

I 2 3 4 5 6
DISTANCE BORON SLEEVES INTO BERYLLIUM (inches)

FIGURE 42. EFFECT OF BORON SLEEVES ON REACTIVITY

 
a7

 

During several runs, uranium foil detector tubes were inserted in the
re-entrant tube to determine the effect of the sleeves on the fissioning
rate in the end duct as described on page 57. The effect of adding this
uranium in this region on the reactivity measurements for the sleeve was
investigated by determining the critical rod position with and without the
detector tube for several sleeve positions. No measurable effect was observed.

Critical Concentration as a Function of Rod Position
and BIU Sleeve Position

The concentration of uranium in the salt mixture was determined frequent-
ly throughout the experiment by performing a chemical analysis on a sample
removed fram the sump. A running inventory of the sump contents was maintained
from which a reasonably accurate uranium concentration could be calculated for
cases when no chemical analysis was available. Thus the variation of critical
concentration with control rod position and with B1O gleeve position can
easily be obtained. By correcting both the rod and BlO sleeve data to the same
temperature, the interrelationship of all three parameters can be shown simul-
taneously as a three-dimensional surface. Figure 13 depicts this surface for
a temperature of 1225°F and Table 7 gives the data obtained at the points where
chemical analyses for uranium concentrations were made.

This surface shows there is little change in the shape of the curve of
concentration versus rod position as the B0 gleeves are inserted. Likewise,
there is little change in thelshape’ ofithe curve of concentration versus sleeve
position as the control rod is inserted. '

Table 7. Control Rod Position as a Function of Uranium
Concentration and BlO Sleeve Position

 

Distance Control Rod is Inserted .into Beryllium (in.)

 

 

 

Uranium
Concentration B1O gieceve Positions
(wt%) 0 in. 1.5 in. 3 in. 5.5 in. ~ 6 in.
10.97 3.66
11.1% 9.1k
11.26 12.47
11.39 | 14.81 |
11.52 , 16 .85 16.70 15.80 11.77
11.64 18.42
11.83 19.93
11.96 21.26
12.09 22.55

12.20 2h,11 23.53 ‘ 20.87
12.48 | 22 .64

 

 
URANIUM CONCENTRATION (weight percent)

1200

11.50

_28_

S
ORNL-LR-DWG 29833

  
 
  
 
    

TEMPERATURE 1255° F

© EXPERIMENTAL
A CALCULATED

MID-PLANE
CONTROL ROD ODISTANCE INSERTED

FIGURE {3. CRITICAL SURFACE FOR PWAR -1 ELEVATED
TEMPERATURE CRITICAL ASSEMBLY

 

 
29

 

The Specific Mass Reactivity Coefficient

 

The specific mass reactivity coefficient was calculated as the ratio
of Am/m, the fractional incresse in the mass of uraniwm in the reactor core
When the uranium concentration is changed, to ak/k, the reactivity caused by
the change in concentration.

The uranium concentration was determined by chemical analysis after each
three additions fram the enricher. For each set of these additions a m/m was
calculated as the ratio of the increase of uranium in the reactor core to the
mass of uranium in the core after the additions according to the relation

fece-f;C fi Ci
Am/m: _—.?.-a_L_L =] - i
¢ Cp Pe Ce

where C; = initial uranium concentration, wt%,

Ce = uranium concentration after three additions of NepWFg to: the
initial concentration, wt%, |

f; = initial fuel mixture density, g/cm3,

fp = fuel mixture density after three additions of Na,UFg to the .
initial concentration, g/cm3, |

m = mass of uranium in the reactor core corresponding to the
uranjum concentration Ce.

Since precise values for the initial final fuel mixture densities
at 1250°F were not availsble, the ratio /1//7 was calculated with the aid
of a rule established by Cohen and Jones® which states that the density of

- any mixture of fluoride salts over the temperature range of 600 to 900°C
is proportional to the density of the mixture at roam temperature calculated
by the formula
N
= M.f

F= =1 J J
N
s (uy/ fe,
J=1
where Mj molecular weight of the jth component fluoride salt,

fj mol fraction of the jth component fluoride salt,
FJ = room temperature density of the jth ccmponent fluoride salt.

 

 

The mean deviation of the ratios of the densities of 15 fluoride mixtures
- measured at elevated temperatures fram the densities calculated by this formula
for room temperature is 3%.

 

2. S. I. Cohen and T. N. Jones, "A Summary of Density Measurements on Molten
Fluoride Mixtures and a Correlation for Predicting Densities of Fluoride
Mixtures,” ORNL-1702 (July 19, 1954).
30

The net reactivity Ak/k caused by each set of three additions was -
calculated in the following way. The rod value in cents at the critical
position for the initial concentration was subtracted from the rod value at
the critical position for the final concentration, that is, the concentration
after three additions from the enricher. Next, a correction was made for the
effect of temperature on reactivity, using the coefficient of -0.473 cents/F
reported on page 32 and the difference in the temperatures at which the two
critical rod positions had been determined. The temperature corrected
difference in the rod value in cents was multiplied by B = 0.0075 to obtain
the reactivity z:k/k caused by the change in concentration. Values of
(am/m)/( Ak/k) calculated for eight increases in the concentration are listed

 

 

in Table 8.
Table 8. Specific Mass Reactivity Coefficients
Room
Uranium Temperature Rod Value
Concen- Fuel at Reactor Temperature
tration Density Critical Temperature Correction (am/m)
(wtd) (g/cmd) Am/m (cents) (°F) (cents) ak/k  (ak k)2
11.13 L.142 - 21.8 1251.0
11.26 L, 1hk 0.0122 45.2 1257.0 +2 .84 0.00192 6.35
11.39 4,146 0.0120 T72.1 1258.0 QL7 0.00200 6.00
11.52 4,148 0.0119 103.1 1255.7 ~1.09 0.00218 5.45
11.64 4,150 0.0109 128.0 1260.5 +2.27 0.00198 5.50 .
11.83 L.152 0.0165 15L4.7 1262 .3 0.85 0.00201 8.21
11.96  L4.155 0.0115 185.2 1253.5 -h.16  0,00192 5.99
12.09 L.157 0.0113 210.9 1258.2 +2,22 0,00204 5.54 .
12.20 4,160 0.0103 244.2 1259.5 0.61 0.00247  4.17

 

a. The mean value of (Am/m)/(Ak/k) calculated as the numerical average of
the eight values listed is 5.90 £ 0.5. The principal source of error in
this value is the uncertainty in the values of the uranium concentrations.
The probable error of *0.5 assigned to (am/m)/(a k/k) reflects a probable
error of *0.01 wt% reported for the results of the chemical analyses.

 

 

The Effect of Temperature on Reactivity

The effect of the reactor temperature on the reactivity was measured
over the temperature range of 1200 to 13500F for a uranium concentration of
11.83 wt%. For this concentration the control rod critical positions were
measured at temperatures 1197.0, 1256.5, 1265.1, 1306.9, and 1355.5°F. These -
temperatures are numerical averages of the temperatures indicated by 16 reactor
thermocouples, identified by asterisks in Fig. 27 of Appendix A. The control
rod value corresponding to the critical positions were read from the curve of )
rod value versus rod position, shown in Fig. 11. A plot of the control rod position
versus reactor temperatures, shown in Fig. 14, is, within the experimental

 

 
ORNL-LR-DWG 298349

 

¢
220}— —0.473 %o

200— —

180—

CONTROL ROD VALUE (cents)

140}—

20—

 

 

 

100 l l
1200 1250 1300 1350

TEMPERATURE F

 

FIGURE 14, EFFECT OF TEMPERATURE ON REACTIVITY

 

 
 

32

errors, a straight line with a slope of -0.473 cents/ofi. Table 9 lists the
control rod positions determined at critical for the five temperatures and
the control rod values corresponding to the critical positions.

Since the control rod sensitivity variation with temperature is smaller
than other experimental errors, the use of the control rod calibration, which
was performed at temperatures near 1250°F, was Justified. These rod sensi-
tivities were determined by measuring positive periods for small control rod
displacements fram the critical positions at 1197.0, 1306.9, and 1355.5°F.
The control rod sensitivities, in cents/in., observed at these temperatures
are included in Table 9. Sensitivities observed for the same positions of
the rod at temperatures near 1250°F are also given for camparison.

Table 9. Variation of Critical Control Rod Position
and Rod Value with Temperature

Uranium Concentration = 11.83% wtd

 

 

Rod Rod Rod Sensitivity Rod Sensitivity

Pemperature Position Value at Temperature at 1250°F

(°F) (in.) (cents) (cents/in.) (cents/in.)

1197.0 21.35 186.7 20.6 20.8

1256.6 19.89 157.6

1265.1 19.68 153.6

1306.9 18.55 132.6 18.0 17.9

1355.5 17.3%2 111.% 15.8 15.9

 

Fuel Importance in End Duct

 

 

The effectiveness of dumping the fuel as a means of shutting down this
experiment depended on the time required to remove the fuel from the first
few inches of the end duct and the reactivity value of this fuel. Fuel drop
times were determined and are reported elsewhere in this report. The
importance of the fuel was investigated by determining the control rod position
for different fuel levels in the end duct. The results could also be used to
estimate the effectiveness of the boron sleeves in suppressing the fissioning
rate in the end duct.

Control rod positions at criticality were found for fuel levels 1, 3,
5, and 7 in. below the fill probe level, which was approximately 1 in. above
the top of the beryllium. Defining the tip of the fill probe as the reference
level of the fuel, the position of an adjustable probe when its tip was at this
reference level was found as follows.

 
e s

33

The salt was raised until it contacted the fill probe. Then the ad-
Justable probe was moved until it Jjust contacted the salt. The salt level
was allowed to drop and the time and order in which the two probe lights
went out was noted. The adjustable probe was then moved in accordence with
these indications, the salt sgain raised and the time between and order of
lighting of the probes. noted. This procedure wes repeated until the probes
were contacted nearly simultaneocusly. Usually a condition was reached where
the order in which the salt contacted and fell off the probee no longer was
consistent, which was taken as an indication-that the probes were essentially
at the same level and the position of the ad justable probe was noted. The
adjustable probe was then inserted the desired distance from this position
and the fuel was raised until it contacted the probe. The helium supply was
then shut off and the fuel height was mesintained during the run by the gas
tightness of the system. The adjustable probe was also used to probe for
the fuel height after most of the runs, and the readings obtained indicated
that during any one run the fuel height dropped less than 1/16 in., the
equivalent of less than a 2% error in the magnitude of the reactivity value
of the fuel displaced.

Because of a possible interaction between fuel height and rod sensitivity,
period measurements were made for each fuel height and the sensitivity of the
rod was calculated. The values cbtained are listed in Tgble 10 along with the
sensitivities obtained for the same rod travel with the fuel at the f£ill probe
level. Since a significant effect was noted, the sensitivity values obtained
for each fuel height were plotted as a function of rod position and the inte-
gral under a smooth curve through these points was camputed to find the loss
in reactivity as the fuel height was lowered. The results are listed in Table 11
and plotted in Fig. 15. Comparing this result with the B1O gleeve data indicates
that inserting the BlY sleeves 6 in. into the beryllium is nearly 80% as
effective from a reactivity standpoint as lowering the fuel level to the same
depth in the beryllium.

Table 10. Effect of Fuel Height on Rod Sensitivity

 

Fuel Height Rod

 

 

Distance Below Position Rod Sensitivity (cents/in.)
Fill Probe at Criticality Fuel Below Fuel at
(in.) (in.) Fill Probe Fill Probe®
0 19.7h 19.4 19.4
1 19.53 19.1 19.0
3 18.61 17.4% . 17.8
5 16.93 14.8 15.%
T 13.%6 8.35 10.6

 

a. These values were determined during the initial calibration of the rod;
see Fig. 10.

 

 

 
 

{cents)

IN REACTIVITY

CHANGE

- 34—

 

 

 

 

—_—
ORNL-LR-DWG 29835
100
[ | I I
sol— ®© BASED ON ROD SENSITIVITY
MEASURED DURING EXPERIMENT
80—
70—
60_._
SO+—
40.__
30—
20—
IOl—
| | | l I |
0 | 2 3 4 5 6 7
DISTANCE OF FUEL BELOW FILL PROBE (inches)
FIGURE 15. EFFECT OF FUEL LEVEL ON REACTIVITY

 

 
 

Pk & L A 35

Teble 1l. Effect of Fuel level on Reactivity

 

 

Fuel Height
Distance Below Loss in
Fill Probe Reactivity
(in.) (cents)
1 3.7
3 20.6
5 48.2
1 91.7

 

Fast-Neutron lLeakage Measurements

 

The relative effectiveness of each of several arrangements of the B1O
sleeves in the upper end duct in the reduction of the leakage of fast neutrons
from that region of the reactor was measured. A fast-neutron scintiliation
counter sensitive only to neutrons with energies above about 1 Mev, and
similar to one developed by H’orny'ak,3 was secured to the top of the reactor
tank for this purpose. The scintillator, a 2-in.-dia, 1/4~in.-thick ZnS-
Lucite disk, was viewed by a DuMont 6292 photomultiplier tube, and both were
mounted in a water-cooled brass can. A water flow rate of about 1 liter/min
was sufficient to hold the photomultiplier temperature near 68°F when the
ambient temperature outside the can was 110°F. Figure 16 showe the location
of the neutron detector. The thermal insulation between the detector and the
reactor assembly was Superex, a calcined diatomaceous silicg product of Johns-
Manville. During all of the leskage measurements the fuel level was at the
tip of the fill probe.

The fast-neutron leakage measurements were made in two serieg, one during
the reactivity evaluation of the B1O sleeves (p. 24) and the other during
the uranium foil exposures described in the following section. The first
series included three arrangements of the sleeves: (l) both sleeves inserted
1.3 in. into the beryllium, (2) both sleeves inserted 3.0 in. into the be-
ryllium, and (3) both sleeves inserted 5.5 in. into the beryllium. Four
sleeve configurations were used in the second series: (l) both sleeves
inserted less than 0.5 in. into the beryllium, (2) the outer sleeve inserted
approximately 6 in. and the inner sleeve inserted less than 0.5 in. into the
beryllium, (3) both sleeves inserted approximately 6 in, into the beryllium,
and (4) the inner sleeve inserted approximately 6 in. and the outer sleeve
ingerted less than 0.5 in. into the beryllium. The effect of the difference
in the fuel concentration and the presence of the uranium foils was within
experimental error. For both series of measurements the counting rate of

 

3. W. F. Hornyak, Rev. Sci. Instr. 23, 264 (1952).

 

 

 
 

_36_

L
ORNL-LR-DWG 29836

 

 

 

 

 

  
   
      
     

 

77 L L L L L L L L L L LA

 

 

 

 

PHOTOMULTIPLIER 1l
TUBE 3 \
«JN-6.25"fiis 1\ /~ FiLL PROBE
HORNYAK | N
SCINTILLATOR \ ]‘ 1
- -
¥ Lo i

 

 

 

 

 

 

..... .{‘,‘. 5‘-“‘6‘1‘- L R o
R RGeS

 
 

 

 

 

   

“CONTROL

 
 

 

T,
) REACTOR TANK)

T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T T T T T T T T T T M W

 

 

 

 

 

 

 

 

 

N K . N
5 :\"' !
6
NNNE § F NN
: N FUEL
BERYLLIUM:

 

FIGURE 46. LOCATION OF FAST NEUTRON LEAKAGE OETECTOR

 
57

 

a BF_ long counter, located on the floor approximately 20 ft fram the reactor,
was used to correct the leakage meagurements for variations in reactor power. ‘

The results are summarized in Table 12, where the neutron scintillation ‘

detector counting rate, corrected for reactor power, is tabulated for each

sleeve arrangement. In Flg. 17 the ratio of the power-corrected counting rate

for each sleeve arrangement to that for both sleeves inserted less than 0.5 in.

into the beryllium has been plotted versus sleeve insertion. These ratios are

also listed in Table 12. The horizontal bars on the data points at 1/4 in.

and 6 in. indicate the uncertainties in the positions of the sleeves for those
points.

Table 12. Fast-Neutron leakage Measurements

 

ZBl'0 Sleeve Positions
(in. Below Top of Beryllium)

 

 

Inner Outer Counting Rate® Ratios of Counting
Sleeve Sleeve (counts/min) RatesP

0.5 (Out) 0.5 (Out): 9870 1.00

1.3 1.3 9280 0.9h

3.0 3.0 8090 0.82

5.5 5.5 6220 0.63

6 (In) 6 (In) 6120 0.62

0.5 (Out) 6 (In) 6910 0.70

6  (In) 0.5 (out) 9080 0.92

 

a. Corrected for reactor power.
b. Ratio of counting rate to that when both sleeves are inserted less than
0.5 in. '

Longitudinal Fission Rate in End Duct

The longitudinal fission rate distribution in the upper end duct for
various positions of the BlO sleeves was determined by exposing enriched
uranium foils, 3/% in. in dismeter and 4 mils thicks, in the fuel annulus.

The foils were distributed between 1/8-in.-thick calcium fluoride spacers

to simulate the fuel mixture. This arrangement resulted in a uranium

density of 0.6 g/cc as compared with O.4 g/cc for the reactor fuel, The
discrepancy is primarily due to the fact that the detectors were designed well
in advance of the experiment when a higher value of fuel concentration was
estimated.

 

 
 

 
 
 

Uy
ORNL-LR-DWG 29837

 

o7

0.6

0.5

0.4

0.3

FAST NEUTRON LEAKAGE (arbitrary units)

0.2

0.1

 

  
   

O B0TH B SLEEVES INSERTED

D INNER SLEEVE ONLY

A OUTER SLEEVE ONLY

  

 

1 | I ]

 

2 3 4 5 6
DISTANCE BORON SLEEVES INTO BERYLLIUM (inches)

FIGURE 17. EFFECT OF B° SLEEVES ON LEAKAGE OF FAST NEUTRONS IN THE REGION

OF THE UPPER END DUCT

_89_
39

Sixty-one foils were arranged in each of six detector tubes. A
representative number of these foils were within a narrow weight range for
counting purposes and the others were selected to give the correct average
density. The detector tubes consisted of 7/8-in.-0D Inconel tubing with a
0.035=-in.=thick wall and 0.076-in.-thick end caps. The calcium fluoride
spacers were made cup-like to isolate the uranium from the Inconel, thereby
preventing intermetallic diffusion at operating temperature. In addition,
aluminum oxide spacers were placed at the ends of the tubes where welds were
made since the stability of calcium fluoride at the welding temperature was
in doubt. The physical arrangement of material in the detector tubes is
shown in Fig. 18.

The detector tubes were introduced into the reactor fuel annulus through
the re-entrant tube (Fig. 1) which was constructed of 1.125-in.-0D Inconel
tubing with a 0.062-in.-thick wall and a 0,062-in,-thick bottom cap. The
exact orientation of the re-entrant tube is not known due to warping and
dimensional changes during the reactor assembly and heating. The last
available orientation placed the tube axis at an angle of O deg 55 min to
the reactor axis with the bottom of the tube closer. to the inner core shell.
The distance from the inner core shell to tube axis at this point was 1.237
in. '

Longitudinal fission rate distribution measurements were made for five
settings of the parameters. (It was assumed that with the bottom of the control
rod below the bottom of the detector tube, changes of rod position would have
little effect on the fission rate distribution in the end duct.) Table 13
gives the reactor conditions during the runs. A total of six runs were made,
five with 20-min exposures at a relative power of 0.05, and one with a
180-min exposure at a relative power of 0.10. The relative fission rates of
the foils from the five 20-min exposures were compared by counting their
normalizers at equal decay times for a period of 45 hr. The long, high-power
run, Vi, was made under essentially the seme effective conditions as run V.
The data are given in Fig. 18 and Table 1k.

Table 13. Conditions of Reactor During Longlitudinal Fission
Rate Traverses in the Upper End Duct

 

Distance Between

 

 

 

Fuel B1O sleeve Positions . Bottam of Control
Concentration (in. Below Top of Beryllium) Rod and Top of
Run (wt%) Toner Sleeve  Outer Sleeve Beryllium Island (in.)
I 10.98 0.5 (Out) 0.5 (out) 6.2
11 11.83 0.5 (Out) 0.5 (Qut) 19.89
111 11.83 6 (In) 0.5 (Out) 18.49
TV 11.83 0.5 (Out) 6 (In) 16.79
Vg 11.83 6 (In) 6 (In) 15.63
Vi * 12.48 6 (In) 6 (In) 20.90

 

* Long exposure.

 
ACTIVITY, ARBITRARY UNITS

 

-40-

ORNL-LR-DWG 29838

 

SERIES BORON SLEEVES CON'LI?EOLWT%OD
ISTA
INNER | OUTER | aemyivim, Woues

out®| out 6.2

 

 

 

ouT 19.89

IN** 18.49

 

 

ouT 16.79

 

v a IN 15.63

 

 

 

 

 

 

 

*0.5-in. INSERTION
*% 6-in. INSERTION

7
/ / / BERYLLIUM REFLECTOR
/ J // 7

INCONEL- U FOILS SPAGERS ityp.)

INCONEL
FUEL

BERYLLIUM ISLAND

CONTROL ROD

 

? 6 5 4 3 2 1 Q
DISTANGE INTO BERYLLIUM ISLAND {inches}

FIGURE I8. LONGITUDINAL FISSION RATE DISTRIBUTION

 
 

Table 14. Longitudinal Fission Rate Distribution in the Upper End Duct

 

Distance from Top of

Island Beryllium®& Relative Activity

 

 

(in.) Run I Run II Run III Run IV Run V, Run V"~
6.23%0 0.483%3 0.4162 0.3806 0.3LL40 0.31%2 0.3142
5.730 0.4504 0.3841 0.3399 0.3109 0.2660 0.2669
5.105 0.4150 0.3487 0.2843 0.2682 0.2141 0.2087
4 . 180 0.3621 0.3150 0.2377 0.2259 0.1558 0.1594
3.855 0.3208 0.2767 0.1999 0.1878 0.1206 0.1261
3,230 0.2762 0.237h 0.1665 0.1561 0.0933 0.0973
2.605 0.2350 0.2905 0.1%22 0.1287 0.0758 0.0786
1.980 0.1948 0.1734 0.1188 0.1068 0.0611 0.0621 &
1.355 0.1559 0.1409 0.0971 0.0831 0.0467 0.0496
0.730 0.1198 0.1106 0.0790 0.0661 0.0%00 0.0422
0.105 0.0881 0.0812 0.0622 0.0477 0.0311 0.0336

-0.520 0.0607 0.0580 0.0478 0.0383 0,0288 0.0302
-1.145 0.047  0.0405 0.0381 0.0363 0.0340 0.0338

 

 

a., Positive direction is down.

% Long exposure.
 

42

The relative activities were obtained by counting the foils in two pairs
of scintillation counters. Each pair was connected to a common RCL AID-P
preamplifier, which, in turn, was connected to an RBL 2204 single=-channel
analyzer through an RCL 2420 linear amplifier. Each scintillation counter
consisted of a 1-1/2 x 3 in. NeI(Tl) crystal mounted on a Dumont 6363 photo-
multiplier tube. These counters were arranged face-to-face to approximate a
hn counter geometry and were assembled in a lead shield 2 in. thick. Both
tubes in each pair were supplied by a single RCL Mark 15 Model 22 high-voltage
supply through a voltage divider so that they could be peaked together. Inte-
gral counts of gamma rays with energies above 0.5 Mev were recorded.

Each foil was counted twice in each counter assembly, with the normalizer
foil from the same tube in the other. The normalizer foil was exposed at the
same time as the counting foils and represented the same position in the reactor
for each set of runs. For each longitudinal run, the foil nearest the bottom
of the detector tube was chosen for the normalizer. Since this method was used,
no decay correction was necessary to determine the ratio of foil activation to
normalizer activation after epplication of background corrections to the ob-~
served counting rates. Counting in this manner also corrects for any slow change
in counter sensitivity with time. Statistics were good to 1%; however, a small
variation in the thickness of the fluoride spacers or change in orientation of
the re-entrant tube would lead to an indeterminate error.

Correctigns for changes in counter sensitivity between runs were made by
counting a Co O standard under the same counting conditions used for the
normalizer foils. Corrections were also made for differences in reactor power
during the various foil exposures. The reactor power was monitored with a
BF3 proportional counter.

The counting rates were also corrected for both the counter background
and the initial activity of the uranium foils. A determination of the counting
rates of ten unexposed foils from the same weight group as those exposed in the
reactor indicated that the variation of this activity between foils was small.
Therefore, the values obtained were averaged and applied to each of the foil
counting rates in the same manner as a background correction. As the background
and average initiael activity were only a small percentage of the gross activity,
the statistics are still within 1%.

As plotted in Fig. 18, the relative activity, R. A., for each foil used to
reasure the *longitudinal fission rate distribution is given by

Cr(t) - Bp - T ooy

R.A, =
Cn(t) - By - 1

where
Cp(t) and Cy(t)

activities of exposed foil and normalizer,
J respectively, observed at time t after exposure,
Bp, By = counter background during counting of foll and
normalizer, respectively,
I = average initial activity of both foil and
normalizer,

 
h3

P = normalization factor for reactor power {from §
BF3 counters),

S = normalizgtion factor for counter sensitivity
(from Co®0 counting rates),

normalization factor to compare aetivities from
different detector tubes (from normalizer foils).

=
i

Radial Fission Rate Distribution

 

 

The radial fission rate distribution was measured at three elevations in
the resctor core and at one in the lower end duct. Uranium foils were used for
detectors. These were placed in detector tubes which were positioned in the
fuel as shown in Fig. 19 and Table 15. '

Table 15. Positions of Radial Detector Tubes

 

 

 

Distance fram Reactor Midplane Angle of
_at Point of Attachment (in.) Core Shell Tube Axis to
Tube In Cold In Hot to Which Reactor Axis
Designetion Assenbly Assembly Attached (deg)
Above Midplane
Dg 11.00 11.11 Outer 76
D), 10.06 10.17 Inner 90
At Midplane
A 0 0 Outer g0
B 0.07 0.07 Inner 90
Below Midplane
D3 11.00 11.11 Outer 104
Do 1.4k 11.56 Inner 90
Dy * 2k, 32 2k .57 Inner 41

 

* ITn lower end duct.

The detector tubes were welded to the core. shells and remained in the reactor
during the entire period of operation. The detector tubes consisted of T/B-in.-
0D Inconel tubing with a 0.035-in.-thick wall and 0.035-in.-thick end caps.

The tube lengths varied with their positions in the assembly. The uranium foils
inside the tubes were separated with calcium fluoride spacers as they were for
the longitudinal fission rate measurement deseribed previously. The counting
procedure was also the same.

 
 

w
©
<
o
Y
q
S
-
O
|
Lt
@
O
=
L
2
T
3
i
@
>-F
=
>
.—
Q
|

DISTANCE FROM REACTOR AXIS

26
TOP

~ENTRAN

20

FIGURE 19.

PWAR-|

 

0 |
DISTANCE FROM REACTOR MID-PLANE (inchesc))

ELEVATED TEMPERATURE CRITICAL
FISSION RATE DISTRIBUTION

ORNL-LR-DWG 29839

O T=00 ON INNER CORE SHELL

V T=025

Q T=050

A T=075

O T=1.00 ON OUTER CORE SHELL

T=FRACTIONAL DISTANCE ACROSS FUEL
ANNULUS FROM INNER CORE SHELL

20 28
BOTTOM

ASSEMBLY

30/83
W5

The radial foil activities are plottefi with respect to the average
relative activity, defined by the equation '

 

J A(r,2) av
Vg -
Agy = Vg
where
V. fuel volume ,
A(r,Zg = relative activity at r,Z,

‘ radial coordinate,
Z = s8axial coordinate.

It

This integration was performed numerically. The fuel annulus was approxi-
mated by a series of 1l cylindrical anpuli (Fig. 19) of height AZ. Each
annulus was Turther divided into a series of concentric shells of thickness ar
and mean radius r. The relative activity for each incremental volume was
interpolated from the relastive activity of the radial foil data. This value
was assumed to be independent of Z over each cylinder. The average relative
activity was then determined fram the summationt

 

14 N
s = A (rij) ry Arl'] AaZg
o =1 g=d
bav = W
EE T em A
where
(r,.,) = relative activity in the jth cylindrical shell of the
iJ
ith section,
ryj = mean radius of the cylindrical shell of the Jjth section,
ALryy = thickness of the ith cylindrical shell of the Jjth section,

AZs; = height of the ith section.

The volume of the fuel region, as determined by this method, is 148.7 liters,
as compared with the measured value of 147.2 liters in the cold assembly. The
activities were then normalized to this average relative activity.

Teble 16 gives the results of all of the radial tubes. Foil positions
in this table are given as the perpendicular distance of the foil from the
inner core shell. This distance, together with the distance of the attachment
point of the detector tubes from the midplane and the angle of the detector
tube axis to the reactor axis listed in Table 15, gives the exact positions of
the 1ndiv1dual foils. The data are plotted in Figs. 20 and 21.

 

4. W. L. Scott, "predicted Power Distribution in ART Core,"” ORNL=CF=55«12-176
(Dec. 16, 1955)

 

 

 
Table 16. Redial Fission Rate Distribution
(Relative to Reactor Average)

 

Distance from
Inner Core Shell
(in.) Relative Activity

 

Tube D5, Above Midplane

 

2.061 0.691
2.546 0.733
3,031 0.825
3.516 1.05k
3.880 1.388
4.123 1.781

 

Tube D), Above Midplane

 

 

 

 

 

 

0.05k 1.08k
0.30h4 0.930
0.679 , 0.798
1.179 0.713
1.679 0.685
2.179 0.711
Tube A, At Midplane
0.174 1.503
0.799 1.090
1.549 0.93%6
2.299 0.863
3,049 0.858
3,67k 0.921
4.299 1.0k2
46Tk 1.174
5.049 1.423
5,299 1.638
5.549 1.987
5.799 2.516
6.049 3.378
Tube B, At Midplane
0.054 1.804
0.30k4 1.491
0.554 1.306
0.804 1.173
1.179 1.050
1.804 0.922

2.429 0.878

 
47

Table 16. {cont.)

-

*‘.[, el

Distance from
Inner Core Shell

 

 

{in,) Relative Activity
| Tube | D3,; Bglow Midpla.ne

2,061 | 0.899

2.546 0.919

3,031 1.0L45

3.516 1.340

3.880 1.764

" ok

4,123 2.325

Tube Dp, Below Midplane

ke » anlienuglh

 

0.054 1.613
0.304 1.327
1,179 0.940
1.679 0.896

24179 0.905

i sttt onnfetisadtateiats

" Tube Dy, Below Midplane

i 4 o . .y e o
o ’ :

0130,4- ' a2

0.679 | 0,322
15544 . 04399
1.929 0,505

2.179 0.631

ittt e P A g - o e s,

a.s Foils destroyed, probably by leak in detector tube.

 

 
 

A
ORNL-LR-DWG 29840

 

 

 

 

 

 

 

 

 

 

 

40 T [ [ | [
O TUBE B-0.69" BELOW MID-PLANE
o TUBE A-AT MID-PLANE
A TUBE D,-11.56" BELOW MID-PLANE
QO TUBE D,-I1.1I" BELOW MID-PLANE (AT CORE SHELL)
A TUBE D,-10./7" ABOVE MID-PLANE L
@ TUBE D,- 111" ABOVE MID-PLANE (AT CORE SHELL!
Led
=
30 < —
&
Q
=
- 1
" o 2 3
O T W i
g n & T
@ W ]
w o o w
x o Q oc
< 1o Q
x O
o & & o
Q = - w
3 z 3 5
x 20 O -
O
|
w > ()
2
& \
-]
‘&J 0]
r A
>
E O e L)
.o \ —
L D Oy
=11 - -
0 l ] l l |
27"
RN 1O 2.0 3.0 4.0 6.0

DISTANGE FROM INNER GORE SHELL (inches)

FIGURE 20. RADIAL FISSION RATE DISTRIBUTION, TUBES A, B, D,—Dg.

7.0
-49 -

 

 

ACTIVITY, RELATIVE TO REACTOR AVERAGE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

el
ORNL ~-LR-DWG 29841
08
0.7
06k
0.5 |
| -
o T
L:E n
« L)
o
g |w
s 3
© ac
T =
Z 3
<
0.3
02}
0.1
0
U FOILS GaF, SPACERS
/gEACTOR AXIS
Al, 04
INCONEL
| | |
| 2 3

DISTANCE FROM INNER CORE SHELL (inches)

FIGURE 21. RADIAL FISSION RATE DISTRIBUTION - TUBE D,
24.57" BELOW MID PLANE AT CORE SHELL

 
50

s g

Counting statistics for these foils were within 1%. Any error introduced
by the numerous normalizations or assumptions will affect the absolute magni- -
tude of the relative activities as plotted but will not affect their relative
values.

Gold Neutron Flux Distribution in the Beryllium Reflector

Gold foils were used to measure the neutron flux distribution in the
beryllium reflector of the remctor at three nominal elevations: the midplane,
8 in. below the midplane, and 16 in, below the midplane. The actual positions
of the foils in both the hat and the cold assembly are given in Table 17. '

Table 17. Longitudinal Positions of Gold Foils in Cold
and Hot Assemblies

 

Distance fram Midplane (in.)

 

 

In Cold _ In Hot
RNominal Position Assembly Assembly
At midplane 0.03 (below) 0.03 (below)
8 in. below midplane 8.0% 8.09
16 in. below midplane 16.03 16.16

 

The gold foils were 5/16 in. in diameter and 2 mils thick. They were held in -
covered beryllium oxide cups and placed in slots in berylljum slabs which were

then placed in the reflector. The foils remained in the reflector throughout

the period of operation. Technical difficulties prevented the comparison of

these bare foils with cadmiumecovered foils (see Appendix A). The final

reactor run was mede nine times as long and a factor of two greater in power

then any other run so that the major part of the activity observed in the

foils would result from this run for which the rod position and fuel concen-

tration were known. In this run the bottom of the control rod was 3.25 in.

above the reactor midplane and the fuel concentration was 12.28 wt% uranium.

When the foils were removed from the reactor, they were found to have
melted and, in some cases, gave the appearance of having vaporized or
diffused into the beryllium oxide. For this reason it was necessary to count
the entire beryllium oxide assembly along with a piece of tape which had been
used to pick up any smsll particles which possibly had fallem out of the cup
into the foil slot in the beryllium slab. In some cases, the additional
material so obtained raised the counting rate by 5%.

The condition of the foils indicated an alloying with some material.
Spectroscopic examination of an exposed foil revealed considerable quanti-
ties of zine. Zinc concentration in unexposed foils was negligible. As a .
halflife determination of the material was in go reement with the
accepted halflife of gold, it is reasonable to ‘&ssume that only gold

 

 
51

 

activity was being counted. Ro conclusions have been drifin as to the source
of the zinc.

The relative actitivies of the gold foils were obtained with the same
counting procedures used for the uranium foils except that gamma rays above
0.33 Mev were accepted. The results are given in Fig. 22 and Tsble 18,

An estimation of the absolute @€rror iIn these meagurements is complicated
by the fact that some material may have been lost. For this reason, the
negative indeterminacy in the results is less than 1%, which represents the
counting statistics, whereas the positive indeterminacy may be large. This
fact was considered when the curves shown in FPig. 22 were drawn.

Teble 18. Radial Gold Neutron Flux Distribution in the
" Beryllium Reflector

 

 

 

Distance from Inner ‘ Relative Activity
Surface of Beryllium O in. Below . 16 in. Below
Reflector (in.) At Midplane Midplane Midplane
0.190 0.473 0.478 0.217
1.125 0.687 0.826 0414
2.125 0.866 0.998 0.412
2.875 0.600
3.125 1.000% 1.008
3.625 0.613%
3.875 0.998 0.951
4,375 0.559
4,625 0.947 0.841
6.375 0.739 0.711
6.500 0.350
T.375 0.595 0.610
74935 0.122
8.875 0.450 0.420
10.375 0.269 0.257
11.710 0.095
13.010 0.063

 

a, This foil used for normalization.

 

7

g™

 
 

 

  
  
  

  
 

  

 

 

 

 

 

 

 

 

 

 

_ag_

 

 

ORNL-LR-DWG 29842
1.0f— 2 A a 0
't
w
ul A Q
€l 5
8l
Q
09— 2
! wl
g »| @ (J )
o = N
g B 3 /X &
d| -
x A A
08— w|
4 =] >
wl wnl a
D | W
w I| @ A
£
07—
® MIDPLANE
- A .
> 0.6/ O A\ 8" BELOW MIDPLANE
g ..
< [7) 18" BELOW MIDPLANE
wl
>
go.s—-
™ QD
@
04t—
03— 16" BELOW MIDPLANE 8" BELOW MIDPLANE MIDPLANE
()
A
02—
CJ
0.tp—
MIDPLANE -
OUTER EDGE OF BERYLLIUM REFLEGCTOR {8"85'—0\" MIDPLANE ‘T \O\
I L ) I 16" BELOW MIDPLANE - : | | | ,
0 | 2 3 4 5 6 7 8 9 10 ( 112 13

DISTANCE INTO BERYLLIUM REFLECTOR (inches )

FIGURE 22. RADIAL NEUTRON FLUX - GOLD FOILS
Fvomihl
Appendix A
ENGINEERING REPORT

In this appendix the various components comprising the critical experi-
ment assembly are described in detail. In addition, some account of the fabri-
cation techniques and problems encountered is given, as well as a simple de-
scription of the operation of the equipment. If greater detall is required,
engineering drawings, circuit diagrame, etc. may be obtained from the Pratt
and Whitney Aircraft Engineering Assembly Specification Index FXR548 and from
Oak Ridge National Laboratory.¥* Drawings of specific subassemblies are indi-
cated by the T numbers shown in Fig. 23, which is a sketch of the entire
assembly.

Description of Assembly Components
Reactor Tank

The reactor tank was constructed of Inconel and served to contain the
entire reactor assembly as shown in Fig. 23. Calrod-type heaters attached
around the entire external envelope of the tank were capable of providing a
maximum of 200 kw for heating, although it was found that about 11.9 kw
sufficed to maintain an average temperature of 1250°F in the assembly.
Risers in various positions on the tank top accommodated thermocouple leads.

The exterior of the reactor tank was insulated with kh-in.-thick blocks
of calcined diatamaceous silica, and the tank assembly was mounted about 5 ft
above floor level on a stainless steel stand.

The reactor tank was pressurized with helium through a gas inlet-outlet
line located on the tank top. A gastight connection between the top of the
reactor tank and the reactor assembly was made by a bellows coupling to allow
for differential thermal expansion. The helium pressure in the tank was held
at about 10 psig or about 4 psig above the maximum pressure expected in the
reactor core. A permanently installed gas line between the reactor tank and
the island region assured equalization of pressure in the two beryllium
regions.

Sump Tank

The sump tank, also constructed of Inconel, had a capacity of 287
liters at operating temperature. It was suspended from a tripod stand
located below the reactor tank (Fig. 23) and connected to the core region
of the reactor through a vertical 3.5-in.-OD Inconel fill and drain line
(0.216-in.-thick wall). Since this line was welded directly to the bottom
of the reactor, and since the sump tank and reactor tank were separately
supported, it was necessary to use a bellows coupling gas seal at the point
of entry of the line into the sump tank to allow for differential expansion.
Risers and fittings on the top of the sump tank provided for a thermocouple

 

 

* Changes incorporated in the design during the course of the experiment
and described in this report are not included on the available engineering
drawings.

53

 

 

 
<<l
ORNL-LR-DWG 29843

UPPER SNOW TRAP

T-291531

 

T-29I1878
T-29i1879

T-291882 — ——7

 

<1 T-291508
A
N

175 KW

W)
o
M
o
o
1
o
-
-

50-3500W, 230V

/1 T-291877

22KW, 230V

F-1172
E-1192 (Ref.}

C-1193
LOWER SNOW TRAP

C-1649 (Ret.)

 

FIGURE 23.PWAR-| ELEVATED TEMPERATURE CRITICAL ASSEMBLY-
OVERALL DIMENSION SKETCH

 

 
55

well, sampling tube, enricher line, level probe, and helium inlet and vent
lines. A large snow trap was placed in the main helium fill and vent line.
The sump tank was wrapped with heaters and insulated to provide a maximum
possible heat input of 72 kw. It was found that temperatures could be
maintained with 3 kw.

Enricher

The enricher was designed as a convenient means for increasing the
uranium content of the fuel mixture in the sump tank. It operated as follows:
a piston, actuated by a hand-operated screw, moved downward into a cylinder
displacing molten NapUFg which flowed over a wier and thence through the
enricher line to the sump tank. The enricher was filled initially to the
level of the wier, and the zero position of the piston was established before
the enricher line was connected to the sump tank. GSubgequently any downward
motion of the piston fram this position, as measured by a.Veeder Root counter
directly geared to the screw, resulted in the transfer of a known quantity of
NagUFg to the sump tank., The calibration indicated that on the average 165 g
of uranium was transferred for each 0.125-in. stroke of the piston as would be
expected fram the calculated value using the piston dimensions. Because of
surface tension effects at the wier, however, the total amount transferred
during any single enriching operation was uncertain by about 40 g. Since
enrichments were typically of the order of 300 g and also since samples were
frequently taken for chemical analysis, no particular problems were encountered
due to this uncertainty.

The liquidus temperature of NaoUFg is 1190CF; therefore, it was necessary
to maintain the enricher at a temperatyre samewhat sbove the rest of the
gsystem. It was decided that 1400F was a convenient operating temperature.
This temperature was maintained by heaters and insulation wrapped around the
cylinder. The maximum heat input which could be provided was 22 kw, but
1. h kv was sufflcient to hold the temperature.

The enricher was suspended by springs and hanger rods from a- bracket
attached to the wall adjacent to the rig as shown in Fig. 6. A means was
-provided to check and adjust the level of the enricher after each enrichment
since proper operation depended upon maintaining the enricher axis vertical

After each enrichment the piston was withdrawn to bring the liquid level
below the weir and reduce the possibility of accidental enrichment. A liquid
level probe was provided in the enricher outlet to indicate when the liquid
was approaching the weir level.

Control Rod Drive

 

The control rod drive, which was supported above the reactor assembly
by a tripod standing on the reactor ‘tank top, consisted of a 1/20-hp electric
motor driving a 10 thread per inch screw. A:traveling nut on this screw
carried a megnet which coupled the control rod to the rod drive. The magnet

 

 
 

56

rode in & sheet metal sleeve designed so that the walls acted as bearing
surfaces for the magnet assembly. A tube extended up through a bearing and
shock absorber at the bottom of this sleeve into the region of the magnet's
travel. A shock absorber, universal Jjoint, and soft iron core was fitted to
the top of this tube. The purpose of the universal joint was to insure that
good contact between the magnet and the soft iron core was always maintained.
At the lower end of the tube a yoke was attached. The control rod was
connected to one end of the yoke. by a wire cable universal joint, and the
other end of the yoke held a weight which served as a counterbalance for the
contraol rod. This entire assembly was mounted in a vertical position and any
interruption of magnet current resulted in the release of the soft iron core
from the magnet, whereupon the entire assembly fell by gravity to its rest
position defined by the position of the shock absorber at the bottom of the
bearing sleeve. Some adjustment in the position of the control rod with
respect to the shock absorber was possible at the wire cable connection to
the. yoke. In addition, means were also provided for small vertical and .
horizontal adjustments of the entire drive unit where it connected to the
tripod, as well as at the base of the tripod legs.

~ Limit switches were mourted so0 that lights on the control panel would
indicate the upper and the equivalent midplane positions of the magnet.
Additional limit switches and associated lights indicated when the control
rod was down and when the magnet and control rod were in contact. There were
other limit switches which stopped the magnet assembly at the upper and lower
end of travel. A synchro transmitter geared to the lead screw made it
possible, through a synchro receiver and Veeder Root counter, to remotely
observe the magnet position, which when the contact light was 1lit also defined
the control rod position. The relation between actual control rod position
and Veeder Root counter reeding was checked immediately prior to each run of
the reactor. This system allowed for driving the rod.at one constant speed,
8.8 in./min. A jog control was also provided which reduced the starting
torque of the motor, allowing for small adjustments in the rod position.

Source Drive

The source drive was mounted adjacent to the control rod drive and was
supported by the same tripod. It consisted of a remotely operated 1/20-hp
motor which drove a 0.250-dia rod through a friction gear. The source was
attached on the lower end of the rod. Guide bearings held the source rod
in position with respect to the drive unit. Adjustment of the source rod
to conform with the axis of the control rod was accamplished by shimming
under the mounting bolts of the source drive unit. Immediately beneath the
drive unit and attached to it, a pig containing Iuacite and boron carbide was
provided to house the source when in the "full out” position. Normally the
source could not be brought into the pig immediately efter removal from the
reactor because of its high temperature. Therefore, provision was made for
the source to stop for cooling at a position between the top of the reactor
and the pig. Limit sw1tches defined the three p031t10ns of the source , i.e.,

ftn

"in} "out," and "full out.

 
ST

Gas System and Muclear Controls

The gas system was composed of three circuits: a filling and mixing
circuit, a helium supply circuit for standby operation, and a moderator
helium suppxy circuit. This system was interlocked with the muclear controls
to provide the following functions:

1. Raise the fuel into the reactor region.

2. Maintsin and indicate the fuel level in the reactor.

3. Mix the fuel automatically.

. Provide interlocks for automatic safety of operatlon.

5. Supply and maintaln an inert helium atmosphere in the moderator
region.

6. Supply and maintain an inert helium atmosphere in the sump tank,
enricher, and reactor core for standby conditions and for filllng
and sampling operatlons.

Figure 24 is an elementary flow diagram of the three gas panels.
Helium was supplied to the gas system from a rack of helium bottles and
manifold through a pressure reducing valve (unnumbered at the extreme right
of Fig. 24). The helium entered the gas panels at approximately 80 psig.
The bourdon gage, PI-8, indicated the supply pressure,and the pressure switch,
PS-5, turned on a red llght when the pressure dropped below 80 psig.

Filling and Mixing Circuit. The gas entered the filling and mixing circuit
at hand valve HV-1l, which was the shutoff wvalve to allow maintenance of the
circuit. After entering at a pressure of 80 psig the gas was then reduced to a
pressure of 20 psig by a pressure control valve, PCV-1l. The pressure was
indicated on PI-3. PCV-2 was a highly accurate Moore regulating walve used to
meintain the pressure in the sump constant to within & small fraction of an inch
of mercury. HV-5 and -6 were used to isolate valve PCV-2 for any necessary
meintenance. Also, since this valve was of the constant bleed type, it was
necessary to isolate both sides of it when the bypess line through HV-L4 was
being used to prevent back flow of radiocactive gases and helium fram the sump
into the control roam through the bleed hole in the PCV-2. HV.T and HV-8 were
the primary manual control valves in the fill line. These two valves were used
in series for safety and finer control considerations. A solenoid valve, SV-5,
was used for automatic shutoff. It was a normally closed valve so that loss of
electric power automatically stopped the flow of filling gas.

Flow indicators FI-1 and -2 were used to indicate the rate at which gas
was flowing into the system and this indicated how rapidly the salt was being
added to the reactor or what the leak rate of the system was when the regulator
valve, PCV-2, wes maintaining the fuel level constant. The flow meters differed
by sbout a factor of 1k in sengitivity. The hand velves HV-9, -13, and -1k were
used to select the flow meter desired and also permitted their removal for
servicing. -

 
 

 

UNCLASSIFIED
ORNL-LR-DWG 29844

VENT

 

 

 

 

 

 

 

 

 

 

 

 

SNOWTRAP, REF NO.2 (SHELL TANK) |

 

 

 

 

_.89_

 

 

 

 

 

 

 

 

   

 

    

NOW TRAP, REF.
NO.1 (FILL 8 DRAIN)

ENRICHER \_,

 

 

 

 

 

 

 

SV = SOLENOID VALUE
HV = HAND VALVE
CV = CHECK VALVE
Pl = PRESSURE INDICATOR
\k_/)/ PS = PRESSURE SWITCH

= PCV: PRESSURE CONTROL VALVE
FI = FLOW INDICATOR

 

GAS CABINET NO.9 (IN TEST ROO

=

GAS CABINET NOS. 7 & 8 (IN CONTROL ROOM}

 

 

 

 

 

 

FIGURE 24. PWAR-I ELEVATED TEMPERATURE CRITICAL ASSEMBLY
INERT GAS SYSTEM
59

 

Pressure in the sump and reactor core was prevented from inadvertently
reaching dangerous values by & high pressure safety switch, PS-1, which
closed SV-5 and vented the system through SV-l. The pressure in the system
was indicated on both PI-1l, a 10-in.-dia bourdon gage, and PI-2, a 35-in.
long mercury moncmeter. In the fill and hold conditions, PI-1 and -2 indicated
the height of the selt in the reactor.

The sump and reactor core were vented with normally open solenoid operafed
valves, SV-1 and -2. SV-2 was closed only in the hold condition. Check valves
in the vent line, CV-l and -2, prevented the back flow of air into the system.

They were &mnnected in parallel to insure vembing. .-~ .

8V-3 and -4 were solenoid velves actuated by pressure switches PS-3 and -2,
respectively, and were used in the automatic mixing operation. PS-3 was set to
open at a pressure which raised the salt somewhat less than halfway up into the
reactor. The electrical circuitrywas arranged (see Fig. 25) so that the actu-
ation of this switch opened SV-3 and closed SV-5, allowing the salt to rapidly
dump into the sump, thereby mixing it. PS-3 closed at slightly above atmos-
pheric pressure, which in turn closed SV-3 and opened SV-5 and again admitted
helium to raise the salt, starting the cycle over again. PS-2 was set to open
at 8 slightly higher pressure than PS-3 and was connected so it opened both
SV-3 and -4 and closed SV-5. It was included as a safety device to dump the
salt in the event PS-3 failed to do so. It also closed at approximately
atmospheric pressure. :

HV-12 was a menual valve which allowed SV-3 and -4 to be bypessed and
was used during the standby condition (see following section) to maintain
the reactor core and sump regions at the same pressure and prevent the salt
from-rising into the reactor core.

The line through HV-11 and CV-3 was provided to allow the filling gas to
flow from the supply through the enricher into the sump. The enriching salt,
NapUFg, has & high affinity for ZrF), vapor which evolved in appreciable
quantities fram the fuel mixture at the operating temperature, and it was
therefore desirable that little or no gas flow from the sump into the enricher.
CV-3 prevented the flow of helium out of the sump through the enricher when the
salt was dumped.

Standby Helium Supply Circuit. During enriching, sampling and sump filling
operations and during other periods when the salt remalned in the sump for an
extended length of time, an inert helium atmosphere was meintained in the gystem.
The helium supply circuit used for this purpose was separate from the one
described above, and since it was used only during non-nuclear operations it was
identified as the "standby" helium supply circuit.

The standby circuit consisted of HV-10, PCV-5, HV-15 and PI-T as shown in
Fig. 24. Two shutoff valves were used so that by opening one, HV-10, before
opening the other, HV-15, it was possible to check the pressure setting of

 
 

 

 

 

UNCLASSIFIED

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
  
   
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

TO TERMINAL BLOCK ORNL-LR-DWG 29845
SL SLN
FRONT OF PANEL P L. ARRANGEMENT ———a O O
DUMP HI LEVEL FILL PROBE ADJUSTABLE SUMP LOW MOD.
R
b L . _ PROBE L PRESS _
i | = = jf = || =+ = ¥ — =
1 Re2 § PS-4 Re! Tn,z Tsan rRe [re2 [ 1 fi R.-| Ryt Tr ! PS.5
fi I PS 2 ROD ETC | 26
—7 1
- n CONTACTSY |
o I ’ R,.ra SCRAM l |
z = — —
& | -—FH_L-:‘: = ’TDIPS-I T 3 1o mop
2 MOD. = ;gfllfi] SCRAM CKT,
3 | PRESS Rt Ly i7aTs83 Toa-u =
T l20a RESET 8 10 PS-3 Tenis 23
x - i,
2 =3 " 1 —12 ' ]
3 | TR —20 22— (M )—— ——- / 27
0 | I 3 4 t |SB-5 |SB-|7 s
> E 14 i6 19 BUZZER
2 l 9 Tse3 i/ SWIToH- ' Posimion THoLD | FiLL [oume | w
) N L A 13 Rel ¥ Ryl N F R L o e L 2 L Posmon Troip | Fiu Joume | mix
O OOOO ¢ b 9 b & o 60 1] @) T e
Ve /s /7 CAL E; ) T, T
| TN\ N TN Tswi Jsva]svs va [ sdal N Ifl =5 TN 5 N\ TN fsve X X X
| (v$-3) | (vs2) & sv-2 X
KEY —- e
| SwiTeH Sv-3 X X X
|
15A SV-5 X X "
ISAMP-2 POLE HIGH LEVEL RUN LIGHT DUMP LIGHT FILL LIGHT ADJUSTABLE LOW MODERATOR KEY RELEASE : - -
TOGGLE SWITCH LIGHT PROBE LIGHT PRESSURE PERMIT LIGHT VALVE POSITlONS
VALVES CLOSED IN POSITIONS INDICATED
By "x"
POSITION HOLD | FILL |[DUMP | MIX
. _
—jaE 1o CONGTLET ¢ CONTACT
DC. TEST SWITCH FOR = P2 1l x X X
SUMP TANK LEVEL O‘”‘O’!l‘gkz
33 3 4 3 X ) L
10A OHFOHKD |4
° 105/115/125-120v-1 § Q BATTERY |w——CONTAINS INTERNAL 5 Sl-o 5 X
POLARITY CHARGER ON/OFF SWITCH
TO BUILDING 500 VA. TRANSFORMER 34 REVERSE {TRICKLE) O’“‘Q_"fi_ G e —d
EMERGENCY SWITCH +| - 7 8 7 X X
SUPPLY I I0A 35 + sl
ADD ADJ. RESISTOR TO OitOHO (8| .
°© o |3 HIGH FILL ADJ. l = ngngE LIMIT CHARGE RATE TO 9 10 g| x X X
15 AMP- 2 POLE LEVEL | PROBE | PROBE = -2 AMPS WITH FULLY o+ ol
PROBE BATTERY CHARGED BATTERY O" o !-9 |'0_ (R Y RSN S
TOGGLE SWITCH REACTOR R.4 no12 {n| x X X |
° OIFHHO (12}
SUMP 13 14 |13; X X |
TANK A " =
iHFHKO |18
15 16 15| X X |
o EMOTE HFOHKO |16 l
GROUNO T0 —w== 7 o8 17| X
BUILDING GROUND~ EMERGENCY SUMP PROBE LIGHTS {OHHEOA Ol 18 4j

FIGURE 25.

PWAR-I

ELEVATED TEMPERATURE

 

 

 

 

 

 

 

DEVELOPMENT _OF SB SWITCH

EVEN NUMBERED CONTACTS NOT REQUIRED AND ARE NOT THEREFORELISTED.
ODD NUMBERED CONTACTS ARE CLOSED IN POSITIONS INDICATED BY "X

CRITICAL ASSEMBLY

ELECTRICAL WIRING DIAGRAM

 

_09_
 

61

PCV-5 with PI~7 before applying any pressure to the sump. This prevented
the inadvertent application of a dangerously high pressure to the sump.

HV-11 was always closed before gas was admitted through this system,
thus insuring that the gas flowed through the enricher into the sump. In
addition, a2 mechanical interlock between HV-15 and -12 made it impossible
to open HV-15 before HV-12 had been opened and also prevented HV-12 from
being closed before HV-15 had been closed, This insured that SV-3 and -k
were bypassed and no difference in pressure between the sump and the reactor
core cauld develop which would raise the salt into the reactor during standby
operation.

Two standby procedures were used. In one the sump was vented’ through
SV-1 and CV-1 and -2 and a continuous bleed of helium through PCV-5 was
maintained. This was used primarily for sampling operations. In the other
procedure SV-1, -2, and .--5 were closed and a constant pressure of 2 or 3
psig vas maintained on the sump by means of PCV-5. The latter procedure was
used for all extended shutdown periods.

| Moderator Helium Supply Circuit. During all operations an inert gas
atmosphere was maintained in the two moderator regions. The circuit which
supplied helium to these regions of the system consisted of HV-2 and PCV-3 and -4
as shown in Fig. 24. Two pressure regulators were used to insure safety by re-
ducing the possibility of applying a dangerously high pressure to the moderator
regions. The pressure was indicated on PI-4 and -5. Pressure switch PS-i wasg
set to open when the pressure in the moderator region fell below & pressure
which was slightly higher than the hydrostatic pressure produced by the normal
height of salt in the reactor core above the bottom plate of the reactor tank.
The opening of PS-4 dumped the fuel and dropped the control rod. This fore-
stalled the creation of the potentially dangerous condition in which the
hydrostatic pressure of the salt in the reactor core was higher than the helium .
gas pressure in the moderator. If a leak in the core shells had occurred under
such a condition, it would have been possible for salt to enter the moderator
region with an attendant addition of a large asmount of positive reactivity; The
control rod was dropped to provide additional safety in the event some salt did

leak into the moderator region. This switch also turned on a flashing light

and rang a bell.
A vent was provided for the moderator regions through HV-3,

Fuel level Indicators. In addition to the indication of the salt level
given by the pressure gauges PI-1 and -2, three probes were provided at the
top of the reactor core region, two in fixed positions and one variasble. The
tip of one of the fixed probes was approximately 1 in. above the-level of the
beryllium island. This probe was called the fill probe and was used to
determine the normal height of the salt for most runs. The salt contacting
the fill probe turned on a light on the gas panel and activated a buzzer. The
buzzer could be silenced by a switch provided in its circuit (see Fig. 25).

 

 
62

The tip of the other fixed probe was 1 in. above the fill probe and was
called the high-level probe. This probe acted as a safety device to prevent
the selt from rising too high in the core region. When the salt contacted
it, solenoid valves were actuated such that the salt was dumped until it dropped
below the fill probe. The valves then returned to the condition they were in
before the high-level probe was contacted.

The adjustable probe was used to determine salt levels other than the
fill probe height and was also used to more accurately determine the salt
level at the fill probe. To do the latter, the adjustable probe was moved
until it was very nearly at the same height as the fill probe and the salt
level was then maintained so it was in contact with one but not the other
probe. In this fashion the level of the salt could be determined and held
to within about 1/64 in. of the levél of the fill probe. Fuel in contact with
the adjustable probe caused lights on both the gas panel in the control room
and the cabinet in the reactor roam to turn on.

A probe was also used in the sump tank to indicate when the salt was in
the sump. This probe, as.well as the other three, was supplied with 110 volts
ac through an isolation transformer from the building emergency power circuit.
Therefore, in the event of a power failure, which should automaticelly dump
the fuel, the sump probe would be supplied with power from the building
emergency motor generator set and could then be used to determine whether the
salt was in the sump. For the minute or so required for the motor generator
set to come up to voltage, or for the case in which it failed campletely, a
storage battery was provided which, by pushing & button, would put 6 volts dc
on the probe. Since a probe in these salts quickly becomes polarized when
direct current is applied to -it, & polarity reverse switch was also included
so that it could be depolarized. .

Connections Between Gas System and Nuclear Controls. The gas system was
connected to the nuclear control circuits in three places. First, as was
mentioned above, loss of the moderator pressure dropped the control rod and
dumped the fuel. Second, the nuclear scram circuit was connected to the gas
system so that initiation of a scram signal also dumped the fuel. And third,
an interlock was provided in the circuit of SV-5 so that when the salt was
in the sump, SV-5 could not be opened to admit gas to raise the salt unless the
control rod and source were both inserted to the midplane of the reactor. This
prevented the salt from being raised into the reactor core when the control rod
or source was in ean ungafe position. After the salt had been raised high enough
in the reactor core so that the probe in the sump no longer was in contact with
the salt, the rod and source could be moved and SV-5 would remain open. This
permitted the Moore regulator, PCV-2, to add gas through SV-5 to maintain the
fuel level during the operation of the reactor.

 

 
 

Nuclear Instruments and Controls. The nuclear instrumentation used for
this experiment was identical with that used for the low-temperaturfi experi-
ments performed at this facility and described in detail elsewhere.
congisted of the following components: two BFxz ionization chambers connected
to vibrating reed electrometers whose output was indicated on linear strip
charts, a BF3 ionization chamber connected to a logarithmic smplifier with a
range of about 6-1/2 decades (this same amplifier also provided & signal to
indicate pile period); a EFx chember connected to & .dc amplifier which was
read on a linear strip char% an anthracene gcintillation counter read on a
linear strip chart; two BF3z proportional counters connected to linear ampli-
fiers and scalers; and a Hornyak scintillator (zinc sulfide in lucite)
connected to a linear amplifier and scaler. The outputs of the two vibrating
reed electrometers and the anthracene scintillator were connected to the scram
circuit. A special control panel, designed and built for the ART Elevated -
Temperature Experiment,l was used again in this experiment with minor
modifications. It contained the switches, lights, Veeder Root counter, and
circuitry for the control of the motion and indication of the position of
the source and control rod and the interconnection of these with the nuclear
scram circuit and the gas system.

 

Snow Trags

Two reaction traps packed with aluminum oxide pellets were installed
in the experiment to remove zirconium fluoride vepor, or "snow', from the
displaced gas during filling and dumping operations. One trap was mounted
in the offgas line fram the top of the reactor; the second trap was installed
in the vent line from the sump tank. The traps were similar except that the
sump tank trap was made from 10-in.-dia pipe while the reactor trap was made
from 8-in,-dia pipe. The sump tank trap was larger because of the higher gas
flow rate through the vent line during fuel dumping procedures.

Each trap consisted of a reaction section and a filter section. The
reaction section, which was 12 in. long, contained 4 to 8 mesh aluminum oxide
pellets which reacted with the zirconium fluoride vapor to form nonvolatile
zirconium oxide and slightly volatile aluminum fluoride. This section was
insulated and heated with heaters to a temperature in excess of 1400°F. The
filter section contained 3-1/2 in. of Demister packing.followed by 1l in. of
Fiberfax packing and another 1 in. of Demister packing. This section was
intended to remove the small amount of aluminum fluoride vapor formed in the
reaction section, as well as any unreacted zirconium fluoride vapor. The
filter section was uninsulated and cooled by natural convection and radiation.

Heating and Temperature Control Circuits

The primary power supply for heating the critical experiment was a
100-KVA transformer feeding a three-phase 4OO-amp disconnect at 460 v.
Fram this disconnect two smaller transformers were supplied. A 75-KVA

 

L. F. T. Bly, J. F. Coneybear et al., "NEPA Critical Experiment Facility,"
NEPA-1769 (April 1951).

 
64

transformer provided the power for a series of variable auto transformers
which controlled the voltage across all heaters except those on the sump
tank. Auto transformers controlling the sump tank heaters were fed from a
25-KVA transformer. 1In the event of a primary power supply failure, a switch
gear was so arranged that the sump tenk variable auto transformers were '
supplied from a 25-KVA capacity diesel standby power system.

A1l the variable auto transformers were mounted in cabinets in the control
room. It was therefore not necessary to enter the test cell to adjust power
to the heaters. The positions of the heater units on the rig are shown in
Fig. 26. |

Thermocouples were located at various positions throughout the assembly
as shown in Fig. 27. The thermocouple leads ran to patch panels where, by
appropriate patching arrangements, they could be made to read on any point
of three 16-point Brown recorders or on two fast Brown recorders. The
16-point recorders were used in each of the three major areas, the reactor,
the sump tank and the enricher. There was duplication in the installation of
thermocouples to allow for failure during the course of the experiment. The
fast Brown recorders permitted continuous reading of any ‘four thermocouples.
They were most frequently used for observing the temperature transient in the
enricher transfer line during additions of enriched fuel and the approach of
the fuel and reactor assembly to thermal equilibrium after filling.

Variable auto transformers were adjusted until all temperatures being
recorded for a given assembly read in a close group at the temperature
desired. The minimum attainsble temperature spreed was typically of the
order of 200F,

Construction gg_Aséembly Components

 

 

As was previously noted, the responsibility for most of the component
construction was assumed by Pratt and Whitney Aircraft; however, the nuclear
control and instrumentation circuitry, the control rod and source drives,
most of the enricher assembly, and the reactor tank support stand were
components which had previously been used at the ORNL Critical Experiments
Facility. In addition, ORNL performed the following tasks: the final
assembly of the detector tubes that contained the foils; manufacture of the
control rod meat and assembly of the rod; drilling of coolant holes in the
larger beryllium island pieces; manufacture of ByC-Cu shields; reassembly of
the enricher barrel; and final setup of the experiment including heaters,
insulation, heating circuits, gas lines and the reactor tank closure welds.

Although the manufacture of most of the assemblies and components
involved standard techniques and methods, fabrication of certain items was
problematical. These problems are discussed below.

 

 
]
AR ORNL-LR-DWG 29846

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[— s
V-9 [ ]
[ ] V-20 V-2iv-22
v-|o{[ J
] ] — | v-23
0 ,\ V-1l
] ¢ / NN
t [ | o [ ]
o ®
» r *
- ®
- | v-17 ||® V=16
o \ f o
[ :
- ® N\
- ®
] _I D VAT
® e
o .
o | 1 N\
- ®
: e > V-l4
- ¢
;] .
L ®
, v-27 | d : >V-13
3 :
V-8 ¢ ¢ .’
A\ /f. ]
v-28 * N / V-18 v
FiLL V-FG + V-19 & 29 [ —— v-12
Vv-24 V-6
V-7 ¢ v-28
v-28
N~
v-30 E‘}v-z
. V-5

 

 

 

<

I

W
— 1
e 0 0 0608 0 00 0 g 0"
® & o & o © o o

 

 

 

 

 

 

 

<A
.

<
1

FIGURE 26. PWAR-1 ELEVATED TEMPERATURE CRITICAL ASSEMBLY
HEATER LOCATIONS

 
e BTG g
=

L
ORNL -LR-DWG 29847

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

     
 

.
RT 30
RT20
RU 20 21
21
0 22
23 RT 10
’ 1§
r
$RrRUIO
i
RU 3| 5
i3
v
BM31™ BM20)] BMIQ BMeq)
“"“"‘fi o] le U 6!
|i. RM
10
12
i3
BLI
s : BL 20,
gEwrio W
L
EP 20
EMAa0O M
41
EP 10O
11~
emio P
i
EL20W
21
ELIt ELIO

 

 

 

 

 

 

 

 

 

 

 

 

 

FIGURE 27. PWAR-! ELEVATED TEMPERATURE CRITICAL
ASSEMBLY — THERMOCOUPLE LOCATIONS

 

 
 

Welding

Throughout the component fabrication period welding was a major problem.
A detailed account of the welding and weld inspection techniques used in the
manufacture of the reactor assembly and the sump tank is contained in CPsS-120*
and CPS-140 which were based on original ORNL Metallurgy Division specifi-
cations PS-1l, -1k, and =18 covering inert gas-shielded tungsten-arc dc welding
for nickel-based alloys. When these specifications, which were very stringent,
were followed, entirely satisfactory welds resulted; however, it was not until
considerable training and practice had been acguired by Canel welding
personnel that good welds were obtained.

Difficulties were also experienced in maintaining the oxygen content of
the inert gas sufficiently low to eliminate the occurrence of oxide inclusions.
The gas used wasfl99.8% pure argon supplied from a shop cascade system through
molecular sieves. The minimum dewpoint of the gas was maintained at -50CF, but
some oxide inclusions continued to occur. This problem was never completely
resolved and it was necessary to relax the weld specifications somewhat to
.allow the presence of not more than one oxide inclusion per inch of weld, no
inclusion to exceed 5% of the weld thickness. |

During fabrication it became clear that the weld shrinkage problem had
been underestimated. In Jjoints where full penetration is required it is,
of course, desirable to begin the weld with an adequate gap between the
gurfaces to be joined. Unfortunately the shrinkage encountered required that
additional fixtures be used, which in some cases resulted in distortion.

Difficulties with Hastelloy X to Inconel joints had been anticipated, but
these joints were no more troublesome than Jjoints between identical metals.
Requirements for filler rod etc. are given in CPS5-120.

Before making the final closure welds on the detector tubes, the tubes,
with their uranium foils and calecium fluoride spacers in place, were
thoroughly outgassed. The closure weld was then made in a dry box containing
helium, thus leaving an atmospheric pressure of helium in the tube. Most
welds were dye checked, leak checked and radiographed.

Fabrication‘gg Core Shells

Developing techniques for fabricating the inner and outer core shells
presented considerable difficulty, particularly for the outer shell. The
cylindrical shell which made up most of the inner core shell was made by

* Canel Process Specifications, Comnecticut Aircraft Muclear Engine Lshoratory,
a division of Pratt and Whitney Aircraft.

 
68

rolling and seam welding a Hastelloy X sheet 0.125 in. in thickness. This
technique did not produce a cylinder that conformed to the desired tolerances
(x0.010 in. on the radius). In addition, repairs to the welded seam produced
considerable localized distortions. It was, therefore, necessary to press a
sizing plug through the shell to help round it and to increase the inside
diameter sufficiently to allow the island beryllium to be put in place. Better,
though more expensive, approaches to this problem would have been to (1) form
the part in a die, (23 roll and seam weld the part to a radius less than that
finally desired and size it in an expanding cluster or (3) produce it by
extrusion.

The technique first attempted in producing the larger components of the
outer shell was hydrospinning. A mandrel was made which conformed to the inside
contour of the liners. The mandrel was mounted on the hydrospin machine and a
rolled, seam-welded conical shell was fixed to it. Attempts were then made to
spin the shell down onto the mandrel,thus producing the final contour. In order
to attain the desired 0.156-in. wall thickness, stock 0.188 in. thick was used
since the metal was expected to move considerably along the mandrel in the
direction of roller motion, thereby reducing the 'shell thlckness ‘during the
spimming process. .

The first problem which arose in this process involved cracking at the
seam weld. This was corrected by rolling the weld bead down until the thickness
at the weld was identical with the parent metal thickness. Circumferential
cracks then began to appear in the region being worked under the roller.  These
were attributed to the very rapid work hardening properties of Hastelloy X and,
therefore, blow torches were installed to elevate the temperature of the
material above the annealing temperature while it was being worked. The fact
that it was impossible to maintain any but a small section of the material
above annealing temperature caused this continuoug annealing method to be
discarded. It then became necessary to dismount the material frequently and
sub ject it to a stress relief cycle.

During the first efforts to form this component, cracks also developed in
regions not being worked under the roller. These cracks always appeared in a
region near the large diameter of the cone that had been contacted by the
mandrel as the spinning operation proceeded from the small diameter to the
large diameter. This was corrected by increasing the apex angle of the cone
until contact was never made between the cone and the mandrel except at
points between the smallest diameter and the region being worked by the roller.

When the above difficulties had been eliminated it became possible to
spin shells free of structural defects; however, the problem of producing
usable core shells was by no means solved. It was found to be virtually
impossible to get good contact at all points between the shell and the mandrel;
therefore, the shells produced did not have the proper contour, the differences
being in many places as great as 0.125 in. on the radius. In addition, no
satisfactory method could be devised to control the shell thickness. The
movement of metal in the direction of the tool motion was small and varied
considerably with distance along the axis of the shell. Consequently, while

 

 
69

the thickness in some areas was reduced, in other areas, particularly those
vhere the required diameter decrease was greatest, the shell wall increased in
thickness. Efforts to fabricate the shells u31ng the hydrospin process were
therefore abandoned.

A second attempt was made to produce the larger components of the outer
shell by die forming in a 1500-ton press. The mandrel intended for use in
the hydrospin process was used as a punch. This method proved quite successful,
although it was impossible to conform to the desired tolerances without first
investigating spring-back and other associated problems and then manufacturing
a new punch and die. Also, since only 0.188-in, stock was available and the
die forming process produces essentially no decrease in wall thickness, the
shell walls were uniformly oversize. It was therefore necessary after die
forming to contour turn the shells down until the wall thickness requirements
were met, Satlsfactory parts were eventually produced by this techniqueo

The smaller parts=of the inner and outer shells of both the lower and
upper duct were constructed by rolling and seam welding them and then either
hydroforming or die forming them. Of these latter processes, die forming
gseemed to give the most satisfactory results. Again some difficulty was
experienced in conforming to tolerances because no time was available for a
full investigation of spring-back, optimum clearances and other phenomena
relevant to the die design. Tables 19 and 20 give the final dimensions of the
inner and outer core shells.

Fabrication QE_Blo Sleeves

Difficulties encountered in fabricating the inner and outer boron sleeves
resulted in changes in the sleeve design during this period. Essentially each
gsleeve consisted of & B1O-filled. annulus formed betwesr two.concentric metal
cylinders. Actually sheets of metal were welded on the cutside and inside of
spacer rings, both at the top and bottom, as shown in the accqmpanylng sketch.

The first attempt to fasbricate these sleeves resulted in a 0.100-in.-

thick boron annulus between 0.019-~in.~thick Hastelloy X walls. The interior
. of the annulus was plated with 1 mil of copper:
" to prevent interaction between the boron and
the nickel component of Hastelloy X. The :
annulus was packed with elemental.Blo powder,
100% of which passed through 100 mesh and 90%
through 200 mesh screen. The annulus was then
outgassed at room temperature and sealed under
vacuun (<1 in. Hg). All welds were dye
checked, leak checked, and radiographed.
During the course of the core shell assembly
the inner sleeve reached a temperature in the
neighborhood of 5000F. At this point it
7' bulged and one of the closure welds failed.

This failure is explained as follows: The

boron powder contained 94% boron; the remain~-

ing 6% was primarily oxygen in the form of

 

 

 

 

 

 
 

 

 

 

 

 

 

 

Jin
{
x| T
21
g
L0
S
g g
4. $ W
Sf ©
N :
Q
Ve
gf?za9fi4fikz
6__
7 =1
S:Z,rzefl495
l-7209/52/

 

70-.

Table 19. Dimensions of Inner Core Shell®™¢

A. Liner Section T 291494

 

Outside Diameter

 

 

(in.)
Station AB ‘ CD
1 8.554 8.473
2 8.542 8.485
3 8.533 8.478
L 8.545 8.430
5 8.543 8.435
6 8.537 8.486
7 8.523% 8.508

 

B. Liner Section T 291495

 

Inside Radii

 

 

Distance from (in. )
Small End

(in.) _Design Actual
0 4.125 . 4,125
1 4.170 4.170
2 4.280 4.300
3 4.520 4.550
4 4,840 4,310
5 5.280 5,340
6 5.875 5.9355
6.5 6.260 6.320
7 6.700 6.730
T 7.050 7

.5 .050

 

a. T_nufibers are Pratt and Whitney drawing
numbers corresponding to the sections being
described. .

b. The wall thickness for each of the three
sections of the inner core shell (liners
T 291494 and T 291495 and stiffener
T 291521) was 0.125 in. + 0.005 in.

c. All dimensions are for room temperature.

 

 
 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

71
Table 20. Dimensions of Outer Core Shell?
A. Dimensions at Specific Statians
Inside Diameter Measured Measured  Maximum
(in.) Deviation Average Deviation
Station in Inside Wall in wWall
(2 in. Mean Diameter? Thickness ‘Thickness
Apart) Design Measured (in.) (in.) (in.)
1 14.850 1k, 765 0.121 0.156 0.010
2 14.850 14.775 0.156 0.010
/= 2 14.850 1L.779 0.156 0.010
| 7 14.850  1k4.795 0.156 0.010
27T 7201491 5  14.850 14.805 0.156 0.010
3 17 435 6 14.850 14.815 0.156 0.010"
g 11237485 g 14.850 14.815 0.156 0.010
‘ 5 L 14.850 14.912 0.012 0.156 0.010
9 15. 058 15.106 0.020 0.166° 0.020
6 10 5 E 15.646 0.01L 0.165 0.016
7 11 16.hgo 0.022 0.160 0.017
12 m 17.485 0.136 | 0.157 0.010
S T 13. h82 18.481 0.064 0.164 0.00k -
g - 1 - 19.46k 0.016, 0.150 0.01
> L CT291492 15 g - 20.270. 0.01 0.151 0.00
/ 16 20.829 0.020 0.15k4 0.005
/7 - 17 21 ou6 21.058 0.010 0.148 0.007
2 —— B BB B B b i
- — 1 20. . © 0. 0. 0.
/3 5 19.90k 19.578  0.200 0.15% 0.012
4 2 21 18 82 18.&8{13; 0.034 0.15L 0.012
15 - 77 in 23 uik 174 0.0%6 0.149 0.011
7 - 2 L1400  16.466 - 0.010 0.152 0.011
. é _ 2 15 576 15.632 0.024 0.15 0.011
17— S 25 E .038 E 0.010 0.15 0.016
18 - 26 850 1 0.152 8.818
_ 3/, ; 2 15. o 2 0.15 .01
. 12 774 7 > 15.713 0.156 0.010
20 - —
21 - ~ |
22 | B. Average Wall Thickness for Each Section
23 - 7 29/492
24 -+ | o Average Wall
25 4 Section Thickness (in.)
‘267 - I.inersd
2 | \/2729/493 T 29148 0.156 + 0.005
’ PO ) QW itE
: ower . + 0.005
723/490 .T 2911+90 0.156 ¥ 0.005
Stiffenersd |
T2 © 0.09% + 0.00
- T 2 hge Eupper; 0.08h ;'O;OOE
T 291492 (lower 0.094 ¥ 0.005
T 291493 0.094 ¥ 0.005
a. All dimensions are for room temperétufe
b. I.e., the difference between two perpendicular

diameters.

c. T numbers are Pratt and Whitney drawing.numbers
corresponding to the sections being described.

d. Gaps between stiffeners and liners = O to 0.020 in.

 
 

T2

boric oxide. As the temperature of the powder increased the boric acid

lost its water of crystallization and the pressure in the annulus increased. .
Gases adsorbed on the boron powder also contributed to this pressure increase

until the elastic modulus of the anmilus walls was exceeded and permanent

deformation occurred. .

Although this problem could have been corrected by outgassing the sleeves
at high temperature, it was decided to rebuild the sleeves to & new design,
since considersble doubt had arisen about the ability of the copper plate on
the interior of the sleeves to completely prevent embrittlement of the
Hastelloy X due to the effect of boron. It was extremely difficult to main-
tain the continuity of the plated surfaces due to scratching during the boron
packing process. The following design changes were therefore made. Type 410
stainless steel, which contains no nickel, was substituted for the Hastelloy X,
and means were provided for continuously controlling the pressure in the
snnulus. Because stainless steel has less strength than Hastelloy X at high
temperatures, it was necessary to increase the wall thickness to 0.062 in.

The boron annulus thickness was accordingly decreased to 0.062 din., some of
which was required to allow for slightly increased clearances between the core
shells and the moderator in the region of sleeve travel.

During assembly of the experiment a vacuum pump was connected to the
boron annuli of the sleeves so that, during leak checks and outgassing of the
moderator regions, the pressure differential across the sleeve walls was never
in the direction of bursting stresses. Iater when the experiment was in
operation and the moderator pressure was the order of 25 psia, pressure in the
sleeves was raised to atmospheric.

Fabrication of ByC-Cu Plates

 

Boron-containing neutron shields in the lower hemisphere of the reactor -
were designed to be a BjyC-Cu cermet clad in stainless steel.* Unfortunately,
efforts to fabricate one of these shields, the largest piece, failed. Because
of this difficulty the shields were redesigned as Inconel  cans containing
boron carbide powder.

The cans of boron carbide were outgassed at roam temperature and the final
closure was made under vacuum (<1 in. Hg). At this time the significant
hydroscopic and adsorption properties of the elemental boron powder was becom-
ing evident. It was decided therefore to check the outgassing properties of
the boron carbide powder with a quick experiment in which a mockup of a boron
carbide shield was made and outgassed at room temperature. The shield was
then heated and the internal pressure noted. Beginning at -28 in. Hg the
pressure increase followed the ideal gas law up to about 140°C, at which point
considerable quantities of gas began evolving. When 200°C was reached, the
pressure was 60 psig. This pressure was then bled to atmospheric and the
heating was continued. The P/T ratio remained approximately constant until
the temperature reached about 500°C, when the evolution of gas was again noted.
The experiment ended when a temperature of about 600°C was reached and a
pressure of 60 psig had again built up. As a result of the experiment plans
to use the boron carbide cans were abolished.

* This BjyC-Cu cermet was developed by the ORNL Metallurgy Division.

 

 
T35

Analysis of the gases coming off the boron carbide powder showed them
to be primarily composed of water vapor with smaller amounts of 50, SOs;, NO,
CO;, CO and hydrocarbons being present. The primary source of evolving gases
appeared to be water of crystallization in the boric oxide impurity of the
boron carbide.

Unfortunately one of the boron carbide shields had been installed in the
core shell assembly before the experiment with the shield mockup was performed
and it was necessary to remove it. The parts originally produced as BjC-Cu
sintered pieces were then re-examined. Although the large piece was badly
warped and cracked, reworking it straightened it sufficiently so that it could
be fitted into position around the lower duct. The reworking caused some
additional cracks, but since structural integrity was of no particular
importance in the neutron shields the sintered pieces were canned and used
despite the cracks. '

The conditions under which the smaller B)C-Cu sintered parts were manu-
factured was as follows: boron carbide was blended with copper powder in the
ratio of 25 vol % boron carbide and 75 vol % copper. This was poured into
stainless steel tubes which were then flattened, evacuated and hot rolled.
The part was then cold formed and machined to bring it within the desired
tolerances. The large plece was sintered first, then enclosed in stainless
steel, evacuated and hot rolled.

Fabrication of Beryllium Parts

The technigues for hot sintering and machining of large beryllium parts
were developed by the Brush Beryllium Company and no fabrication problems
were encountered due to the processes used. However, the excessive machining
time required to produce each beryllium part delayed delivery dates long
beyond the original schedule. This holdup was somewhat alleviated by supply-
ing Brush with extra drill presses required to drill approximately 800 coolant
holes in each h-in.-thick reflector ring.

A small percentage of the coolant holes were chipped at the bottom
surface of the rings. This apparently occurred as the drill bit broke through
during the drilling operation. For the purpose of the critical experiment
this defect was of no consequence.

The deep hole drilling in the two long island pieces was performed at
the Y-12 Plant. The deviation from the designed path of the coolant holes
in these pieces was the order of 0.003 in. per inch of hole length. This de-
parture from design tolerance was also of no consequence in the experiment.

Fabrication of Gold Foils

As previously noted, gold foils used for activation measurements during
the experiment were placed in slots provided for them at various radial
positions in the beryllium reflector. Since an investigation of the compati-
bility of gold and beryllium at high temperature showed a strong tendency for
the two metals to alloy, an attempt was made to find a suitable covering for
the gold foils that was also compatible with the beryllium. Tests with

 
Th

ceramic wafers of both beryllium oxide and aluminum oxide in contact with
gold on one side and beryllium on the other showed no detectable degeneration
of the gold foils at high temperatures (14OOOF) in an inert atmosphere. On
the basis of these tests it was decided to use hot-pressed beryllium oxide as
the gold foil holders.

Despite the above results, it was found when the gold foils were removed
at the end of the experiment that considerable physical change in the foils
had occurred. ©Spectral analysis showed the presence of sufficient quantities
of zinc in the gold to form a low melting alloy. The source of this contami-
nant is unknown but it was established that the original foils and the ceramic
containers were zinc free.

Some effort was expended in attempting to find a cadmium-containing
cover for the foils in the hope that a comparison of the total activation with
the activation by neutrons above the cadmium cutoff could be made. The only
materials which had suitable physical properties and were available were
cadmium oxide and cadmium silicate. A series of experiments were run to check
the compatibility of these materials with gold and beryllium and the results
were negative. Because time was not available to lsunch a thorough investi-
gation of this problem, the effort was dropped.

Fabrication of Control Rod

 

The control rod meat was made by the ORNL Metallurgy Division Ceramics
Iaboratory by pressing and sintering a blend of rare earth oxides and nickel
powder into the form of cylindrical annmuli about 2.3 in. long. Sixteen of
these cermet pieces were then machined and used to make up the poison section
of the control rod. The resultlng cylindrical annulus was contained between
two Inconel tubes.

The first rod that was fabricated had a constent outside diameter;
however, it was found that in order to maintain the desired clearance between
the rod and its thimble (0.062 in.) straightness requirements on the rod
weldment, thimble weldment and upper thimble extension would be impossible to
meet without a major alteration in fabrication technique. It was also found
that when the cold rod entered the hot thimble it tended to bend at points of
contact with the thimble wall because of the more rapid heating of the rod at
these points. This resulted in a blnding of the rod in the thimble.

In an attempt to eliminate the binding problem, the design of the rod was
changed. The outside tube of the rod was removed between the top of the poison
section and the rod lifting flange and replaced by a tube with a smaller
outside diameter (see Fig. 28). With this change it was only necessary for
the 0.062-in. clearance requirement to be met over 38 in. of the rod length
rather than over its entire length. Considerably more curvature of the
thimble and the rod could then be tolerated and no further problems were
experienced with control rod motion.

 

 
 

ORNL-LR-DWG 29848

 

0.215" "
0.035 . ,
2-1/4"long x |6 pieces

e il ottt ol P I P B P 5 R R B B B P R P

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L et et it et ettt il et

 

 

 

[
l
|

 

 

 

 
 

 

 

 

ol |l

 

l

 

1.93"

 

 

 

 

375("+.005"

 

2.185/'t 005"

 

 

 

 

 

\\\ ' / ////

\
\
flireer |

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

   
  
 

 
     
  

 

   

 

 

 

 

 

    

e e = =
e SN
”'I ~—1/2' CERMET COMPOSITION : 70 % Ni
30 % LINDSAY MIX

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7/8" 325" M 38

 

 

 

 

FIGURE 28. PWAR-I ELEVATED TEMPERATURE CRITICAL ASSEMBLY CONTROL ROD

 

 

 

 
76

Fuel Manufacture

 

/
P

The fuel was manufactured by the ORNL Materials Chemistry Division. The
procedure used was as follows: The NapUFg and NeF-ZrF}, . were made separately.
In each case, a low hafnium content mixture of the raw fluorides was prfpared.
The mixture was heated to 1500°F in ai hydrogen fluoride atmosphere ... Tree .
processing steps then were performed. In the first step reduction of U
sulphates and oxides to sulphides and lower order oxides was accomplish@l with
hydrogen gas. Next hydrogen fluoride was used to convert the reduced. products
to fluarides and drive off hydrogen sulfide. Finally hydrogen gas was again
used to reduce the metallic impurities to base metals, therepy releasing
hydrogen fluoride. . '

A more detailed account of the fuel manufacturlng processes is presented
elsewhere.

Material Failures

 

The ma jor material failure which occurred during the course of the experi-
ment was & leak in the fabricated 6-in. schedule 40 Inconel pipe cap at the
bottom of the sump tank. When the final assembly was complete and all checks
were made the sump tank was filled with barren salt (NaF-ZrFj). A few hours
later a heater failure occurred on the bottom of the sump tank at the pipe cap.
This heater was disconnected from the heater power circuits and, since no other
symptoms of a failure were evident, operational checks were begun. Among other
things these checks included testing the mixing operation and raising the salt
into_the reactor core to check the level probes. About a day after the filling
operation had been completed & strang smell of hydrogen fluoride was noticed -
in the test cell. Examination of the sump tank bottom revealed a salt leak.
The barren salt was immediately removed from the sump tank and power to the
heaters was cut off. Removal of the insuletion disclosed extensive leaking at -
the 6-in. pipe cap. Cracks in the walls of the pipe cap extended for about
an inch vertically up to but not into the weld that Jjoined the cap to the
bottom of the tank (see Fig. 29). Extensive fluoride attack had occurred all
over the tank bottom and even up along the tank walls. All insulation and
heaters were removed and the tank was completely cleaned with abrasive wheels.
The 6-in. cap was removed and an 8-in. forged cap was welded in position to
replace it. After the heaters and insulation had been Teinstalled, the tank
was again ready for service.

Re-examination of the X rays of the weld of the 6-in. cap to the sump tank
disclosed no defects either in the weld or in the parent metal, although the
X -ray sensitivity for the parent metal was only the order of 5% Dye checks
made during the fabrication of the sump tank also disclosed no defects. An
additional pipe cap which had been made from the sgme piece of Inconel bar
stock as that used in the assembly was also die checked and radiographed, but
again no defects were apparent.

 

5. E. L. Youngblood et al., "Aircraft Nuclear Propu131on Fluoride Fuel ’
Preparation Facility," ORNL-CF-54-6-126 (June 1954). |

 

 
 

_??_

O
td
i
W
0
<
-
O
=
=

 

 

4 UNCLASSIFIED

 

 

 

¥ = 2400

 

(&) Inside Crack.

Fig. 29. Sump Tank Pipe Cap Failure — (@) Outside Crack

 
 

78

In investigating the failure of the cap the following points were
noted:

1. During fabrication, the pipe cap weld had been subjected to approxi-
mately 12 repairs. There had been no subsequent stress relief.

2. The pipe cap had been machined from Inconel bar stock, which had been
substituted for the originally specified wrought Inconel due to a
delay in shipment. Defects in rolled Inconel plate and bar are
frequently detectable only in sonic tests, and the cap could therefore
have been defective prior to installation.

3, The pipe cap wall thickness was 0.280 in. This was Jjoined to the
bottom plate of the sump tank which was 1-1/2 in. thick. During
initial heat-up, unequal heating of these two parts could have caused
stresses due to differential expansion.

4. The tenk temperature during the initial fill had been 1250°F, while
the temperature of the entering salt during the fill had been 1300°F.
There had also been a temperature differential between the reactor
core and the sump tank of no greater than S50°F during the aforemention-
ed operational checks. Thermal shock from these sources would appear
to be negligible.

5. It is the general opinion that the heater failure at the pipe cap
occurred after the leak had begun, although this could not be
definitely established. The failure could have been due to any or
all of items (1), (2), and (3) above.

Another gmall leak developed during the course of the experiment at
& Swagelok fitting in the sump tank sampling line. The first indication of
this legk was the inability to hold the fuel level during a refilling operation.
During this operation some fuel was always forced up into the sampling line
(see Fig.23), which extended down into the sump tank, until the fuel level in
the sump tank dropped below the bottom of the line. The leak in the line above
the tank was suspected when a loss of the gas pressure in the sump became
apparent. Investigation showed that the leak did exist but fortunately only a
few hundred cubic centimeters of fuel had been lost.

This amount of fuel was sufficient to cause considerable fluoride attack
to the sump tank top. After removing the insulation and investigating the
damage, it was deemed necessary to remove the heaters on the top of the sump
tank from the heating circuits. It was possible to maintain the sump tank
temperatures without these heaters so the experiment was comtinued with no
replacements being made.

Upon removal of the uranium fission rate detector foils from the detector
tube Dy, it was found that a few of the foils nearest the inner core shell were
damaged. This, plus discoloration of the inside of the tube body and the -
calcium fluoride spacers in the same region, indicated a very minor leak in the
detector tube, although no defects had been detected when the tubes were leak
and dye checked prior to installation. It was necessary to machine through

 

 
T9

   

the weld at the tube end in order to remove the foils; therefore, the
probable source of the failure was destroyed prior to its discovery and no
investigation of the origin of the leak was possible.

Toward the end of the experimental program two of the 1/8-in. Inconel
actuating pins on the outer boron sleeve failed at the weld which Joined them
to the upper support ring of the sleeve. Subsequent movement of the sleeve was
impossible and it remained in the "in" position for the remainder of the experi-
ment. This failure was believed to be due to too small a diameter of the
actuating pins, flextre of the pins when in use and the tendency of the sleeve
to cock and bind while being moved. When the reactor assembly was disassembled
these actuating pins:were observed to be badly bent at their lower ends.

Considerable difficulty was experienced with sheathed thermocouple wires
used in the test. These commercial thermocouples consist of chromel and
alumel wires surrounded by a packed powder magnesis insulstor, the whole con-
tained in an Inconel sheath. A marked tendency for these thermocouples to
short and ground out was observed. Investigation proved the alumina to be
very loosely packed in certain regions and sometimes even completely absent.
Wherever possible, beaded thermocouple wires were used.

Typical Sequence of gggratiofi

Operation of the assembly was performed by & crew of four persons for
each shift: a crew chief, a gas panel operator; a nuclear controls operator,
and a recorder. The steps included in the operation consisted of enriching,
mixing, and sampling the fuel, checking the operation of the instruments,
filling the core, bringing the system to critical, maintaining the required
nuclear power level, and finally shutting down the system. These steps are
descrlbed below.

Enriching

In order for the enricher to function properly it was necessgary that its
axis be vertical. Two bubble levels, at right angles to each other, were
permanently mounted on the enricher lead screw hou51ng so that the level of
the enricher could be checked prior to each enrichment. Because of changes
In the ambient temperature in the test cell plus small changes in the temper-
&ture of the experiment components, with resulting changes in thermal ex-
pansion, it was frequently necessary to make level corrections. This.was ac-
complished by adjusting the support brackets which ran from the enricher
barrel to the adjacent wall of the test cell.

When the enricher was not in use its plston was backed off about three
turns (3/8 in.) to ensure that the NagUFg level was well below the wier and
thus to preclude the possibility of any accidental enrichment. The first
step in enriching was therefore to raise the level of the NasUFg to the wier.

 

 
 

80

The required position of the piston for raising the NaosUFg level was de- *
termined by noting the maximum reading of the enricher Veeder Root counter
during the previous enrichment. The approach of the NaoUFg to the wier

level could be noted by observing the resistance reading on a Simpson meter
which was connected to the enricher outlet liquid level probe. When this
probe was contacted, the Na,UFg level was about 0.16 in. below the wier. When
the wier level had been reached, any desired amount of uranium could be added
to the sump tank by continuing to lower the piston into the enricher barrel.
One turn of the hand wheel was equivalent to 165 g of uranium.

Since the enricher temperature was 1400°F and all other components in
the experiment were maintained at about 12509%F, it was possible to observe
the flow of the enriching salt to the sump tank by noting the temperatures
recorded by thermocouples attached to the enricher transfer line. After each
enrichment the hand wheel was backed off about three turns as previously noted.

During each enrichment, Veeder Root readings were made as follows:
(1) the reading found prior to the enrichment, %2) the reading at which the
outlet liquid level probe was contacted, (3) the reading to which the piston
was lowered, (4) the number of turns added, and (5) the reading to which the
piston was returned after the enrichment. Because of the importance of this
operation readings 3 and 5 were checked by a second operator.

Mixing

A description of the mixing operation is given with the description of
the inert gas system (p.57). Analysis of samples during the first phase of
the experiment verified that 15 mixing cycles were sufficient to assure uniform
mixing. The helium supply pressure was adjusted to give a mix cycle time of
gbout 2 min. As the fuel was being raised for the first mix cycle of each
series, response of the period meter was observed. The apparent period was
never allowed to become less than 30 sec; therefore, with the higher uranium
concentrations toward the end of the experiment, it was necessary to increase
the mix cycle time somewhat. It will be recognized, of course, that the
positive period observed during fuel mixing and raising operations did not
necessarily represent positive reactivity of the system.

Also, during the first mix cycle, operation of the mix pressure switches
and associated circuitry was checked for proper functioning. After initial
ad justment of pressure trip levels no difficulties were encountered through-
out the experiment. ‘

Sampling

Fuel samples were taken for chemical analysis after every third enrichment -
following attainment of the first critical concentration. The sampling pro-
cedure was as follows: First the sump tank was vented. A graphite-lined,
flanged pot containing a smsller graphite-lined stainlesgss steel cup was then
connected to the sampling line. The sampling line was then evacuated and
heated to greater than 14OOSF by electric heaters and, on uninsulated parts

 

 
81

of the line, by a blowtorch. The vacuum applied to the flanged pot caused
fuel to be drawn from the sump tank into the pot, spilling first into the
stainless steel cup, then overflowing into the pot proper. The beginning of
this overflow was detected with a liquid level probe, and immediately after
it began the pot was pressurized and fuel remaining in the sample line was
blown back into the sump tank. The pot was then disconnected and taken to
an ORNL Analytical Chemistry Division laboratory for sample analysis. (This
lsboratory and the experiment were located in . ‘the same building. . |

Instrument Checks

 

In general, complete instrument checks were not made prior to each run,
but they were made at least once during each 8-hr shift. These consisted of
checking the response of the BF3 proportional and Hornyak button counters and
checking the nuclear scram on two continmuously reading BF; ionization chambers
and one gamma-ray sensitive channel. Each channel was chécked separately and
the control rod was allowed to drop. The fuel dump was checked on at least
one channel. Response of the log N period meter and one dc amplifier channel
not connected to the scram circuits was also checked. All checks were made
using a Po-Be neutron source or a radium gamma-ray source.

Tmmediately before each run, the relation of the Veeder Root counter
reading on the control rod panel and the physical position of the rod was
established. The rod was brought down to the midplane limit switch and the
Veeder Root reading was again recorded. At the same time the distance between
the control rod flange and the top of the rod thimble extension was physically
measured and recorded.

Filling the Reactor Core

 

With the fuel in the sump tank, the control rod was brought to midplane

and the source put in the "in" position. As previously noted, interlocks were
provided which required this before the helium pressure could'be placed on the
sump tank., The fuel was then raised, using PCV-2 (see Fig. 24) as the maximum
pregsure control valve and FI-1 to indlcate the rate of gas addition, to the
level of the fill probe, noting carefully the response of the period and power
level meters as a guide to the rate of rise. This system of filling gave an
almost idealized startup condition since the rate of gas addition varied in-
versely with sump pressure and the sump pressure increase rate decreased with
increased sump pressure. This rate decrease was a result of the increased gas
volume in the sump as the sump pressure increased. Because the temperature of
the sump tank was not usually identical to the temperature of the reactor core,
it was necessary to keep the fuel within the core for about 45 min prior to
each critical run to allow for thermal equilibration. If this were not done,
temperature transients would disturb the reactivity measurements being made.
A thermocouple in the variable probe which read fuel temperature in the top of
the core and a thermocouple on the outer core shell were continucusly recorded
by a fast Brown recorder located on the gas panel. By this means the approach
to thermal equilibrium could be observed.

Temperature Reedings

 

As noted under "Experimental Results," average temperatures for each run

 

 

 
82

were determined by reading a series of selected thermocouples in the reactor
assembly. These readings for each run were usually taken shortly after the
neutron level was reached and shortly before the control rod position was
recorded in an attempt to obtain a mean of the slow transient temperatures.

Approach to Critical (Routine)

 

When all of the sbove steps had been accomplished the experiment was
ready for operation. The estimated critical position of the control rod was
noted by the operator and withdrawal of the rod from the midplane began. The
rate of rod removal was determined by the period meter which was never allowed
to be below 30 sec. When the neutron level was sufficiently high, removal of
the source began. The criterion for complete source removal was the low
response of the neutron level to source motion. With the source out, withe-
drawal of the rod continued until the rod was at the critical position for the
last enriclment. The power level was then allowed to increase on a period
defined by this rod position. Increments of uranium added were always such
as to make this period the order of 100 sec during a typical run. When the
log N meter indicated the desired level, the rod was moved in until reactor
power was level and the period infinite. The change in power level during
the constant period was usually the order of two decades. Measurement of this
period determined the reactivity difference for the rod positions. The
second measurement was made by reducing the neutron level for a short interval
and then repeating the positive period step.

Maintaining the Power lLevel

 

Since automatic means were not provided, the power level was maintained
by the operator. A tendency of the system to drift in level was attributed
to slow thermal transients and convection currents in the reactor assembly and
fuel and fuel level drift. Thus slow and unpredictable reactivity changes
required frequent minor adjustments of the control rod position.

Two means were provided for maintaining fuel level: autamatic regulation
of the pressure through PCV-2 (see Fig. 24) could be used, or the helium
supply could be closed off. Of these, the latter means was more frequently
used but because of the finite helium leak rate of the system it was necessary
to occasionally read just the fuel level during a run. Whenever the helium
leak rate became excessive it was necessary to remove the accumulation of
ZrF), from the valve seats of SV-3 and SV-bk (Fig. 24).

shutdown

The system was shut down by dumping the fuel and returning the control
rod to the midplane. Immediately after shutdown the source was moved to the
"in" position. After a short wait the test cell was entered and the gas
system put in standby operation.

 

 
Appendix B

COMPOSITIONS AND WEIGHTS OF REACTOR MATERTALS

Teble 21. Camposition of BIO Pieces

 

Material wtd
Total Boron g4
Iron 0.23
Oxygen ~s6

 

Table 22, Composltion of Control Rod Cermet

 

Material | wt%
Nickel 70
Lindsay Mix® 30
(Rare Earth Oxides)

Sm (b) 63.8
cd (v) 26.3
Dy (b) L.,8
Nd (b) 0.9

h,2

Misc. (b)

 

a. The europium was removed fram the mixture.
b. In the form of an oxide; the weight percent
quoted includes the oxygen.

Table 23. Composition of Beryllium Oxide Foil Holders

 

Material wth  Material wtb
Al 0.2 Mg ~ 0.05
B 0.002 Mn = 0.02
Ba. < 0.02 Mo < 0.02
Ca < 0.1 Na, < 0,01
Cu < 0,05 Ni < 0,05
Cb < 0.1 Pb < 0.1
cd < 0,001 S1 0.1
Co < 0.05 Sn < 0.05
Cr < 0,05 T4 < 0.02
Fe < 0.02 v < 0.02
K < 0.01 Zr <0.l
Ii < 0.01 BeO Remainder

 

83

 

 
Table 24. Composition and Weights of Beryllium Reflector and Island Parts

 

 

 

 

 

 

 

 

» Compositiofib : a
a P and W wth, - ppm Density Weight
Piece Dwg. No. Be BeO Fe Al Mo L1 Co N cd B (g/cc) (kg)
Reflector

A T 291496 98.29 1.72 0.15 0.025 90 ¢0.3 5 190 20.2 1.1 1.843 47.2

B T 291497  99.20 1.18 0.12 0.041 80 <0.3 5-10 150 < 0.2 0.8 1.840 64.8

C T 291498 98.85° 1.38  0.15 0.03%8 80 <0.3 3 190 £ 0.2 0.1 1.840 68.2

D T 291499 98.90 1.54 0.16 0.057 70 <0.3 20 400 ~0.2 0.8 1.84  118.4

E T 291500 98.50 1.62 0.15 0.019 120 0.3 3 260 £ 0.2 0.2 1.853 139.2

F T 291501 98.90 1.46 0.10 0.025 120 0.3 b 160 40.2 1.3 1.852 167.8

G T 291502 98.90 1.50 0.14 0.020 130 <«0.3 4 200 £ 0.2 1.3 1.850 175.8

H T 291503 98.60 1.81 0.12 0.026 130 <0.3 L 160 <0.2 1.8 1.849 172.5

I T 291504 98.60 1.78 0.1k 0.013 130 <«0.3 L 180 <0.2 0.6 1.849 168.6

J T 291505 98.60 1.61 0.15 0.019 120 «0.3 3 260 . 0.2 2.0 1.853  140.6 -éz

K T 291506 98.90 1.39 0.1k 0.0k41 80 £0.3 6 230 0.2 1.1 1.856 118.8 '

L T 291507 98.67 1.46 0.15 0.028 100 £0.3 L 210 -0.2 c 1.8k45 69.3

M T 291508 98.40 1.48 0.14 0.046 80 £0.3 1 160 <0.2 1.1 1.847  66.9

TOTAL 1518.1
Island

1 T 291509 98.38 1.70 0.12 0.0k41 80 £0.3 1 90 <0.2 0.6 1.860 13.46

2 T 291510 98.30 1.81 0.15 0.035 90 ~0.3 2 2L0  «£0.2 c 1.854 5.83

3 T 291511 98.80 1.40 0.15 0.025 90 «£0.3 5 270 £0.2 1.4 1.860 26.83

L T 291513 98.40 1.6k 0.13 0.031 80 £0.3 2 150 Z0.2 ¢ 1.857 20.33

TOTAL 66.45

 

 

o

Refer to Fig. 23, p. 5h.

Four samples were checked for rare earths; the results all showed that Gd, Eu, Sm, and Dy constituted
£0.0005% of the samples.

Not detectable.
d. Meagured weight.
Table 25.

Composition and Weights of Core Bhells

 

 

 

 

 

 

 

 

Composition
P and W wtdb ppmP Weight
Piece® Dwg. No. Mo Fe Cr Co Mn W Ni In B Dy Ca Eu Sm Gd (kg)
Outer Core Shell
Top liner T291489 9.41 17.2 23.3% 1.45 0,70 bal. c 0.001-0.0001 c 15.12
Middle linerd T291488
Upper, A 8.2 20.8 22.9 1.71  0.71 bal. - - 5 - 5 5 5 22.53
Upper, B T.4 17.6 23,0 1.69 0.9 - bal. - - 5 - 5 5 5
Lower, C 8.5 19.5 18.4 1.27 0.58 bal. - - 5 - 5 5 5 21.95
Lower, D 8.5 17.3 22.5 1.55 0.48 - bal. - - 5 - 5 5 5
Top stiffener T291491 7.91 17.0 21.7 1.7%  0.67 - bal. c 0.001-0,0001 c 9.13
Middle stiffener T291492 -
Upper 8.55 17.8 20.1 - 0.92 - bal. - - 5 - 5 5 5 7.4k e
Lower 6.76 16.8 21.6 1.2 0.61 bal. - - 5 - 5 5 5 7.42
Bottom liner T291490 10.2 17.1  22.5 1.19 0.5 - bal. c 0.001-0.0001 c 10.81
Bottom stiffener T291493 8.69 17.2 22.9 1.21  0.56 - bal. c 0.001-0.0001 c .
Total  9L.LO
Inner Core Shell
Top liner T2914 9k 7.50 17.3 24.5 1.61 0.72 - bal. - - 5 - 5 5 5 23,04
Bottom liner T291495 7.66 16.9 23.3 r.12  1.21 - bal. - - 5 - 5 5 5 8L
Bottom stiffener T291521 7.50 15.0 2L4.5 2.1k 0.62 - bal. c 0.001-0.0001 c 5.
Total 28.88

 

a. Refer to Tables 19 and 20, pp. 70 and T71.

b. These values are lower limits of the analytical method and the actual contents are less than the amounts listed.
c. Not detectable.

d. Sample A was taken near the top of the upper section of the middle liner, while Semple B was taken near the bottom of the upper section.
Similarly, samples C and D were taken near the top and bottom, respectively, of the lower section of the middle liner.

 

 
Appendix C

LIST OF FIGURES

Figure Ho.
1. Reactor Assembly - Major Components . . . . .
2. Core Shell Agsembly in Place . . .+ . ¢« =«

1k,
15.
16.

17.

18.
19.

20.

Partial Reflector Assembly . . .« . « « =«
Reactor Assembly Complete . . « « + . .

Repetor Tenk in Place with Heaters . . . . .
Assembly Complete .« . e e e s s » -. .

A Comparison of the Power Reactor Design (PWAR-1)

the Elevated Temperature Critical Assembly . .

, Reactor Assembly - Dimensions . . . « ¢« =«

Reciprocal Multiplication . . « + « + « .
Control Rod Sensitivity « . +« + + + .. .
Control Rod Evaluation « . + +» « « « =«
Effect of Bl0 Sleeves on Reactivity . . +» « .

Critical Surface for PWAR-1l, Elevated Temperature
Critical Assembly O

Effect of Temperature on Reactivity . . . .
Effect of Fuel lLevel on Reactivity . . « .+ &
Location of Fast-Neutron leakage Detector . . .

Effect of BlO Sleeves on leakage of Fast Neutrons
the Region of the Upper End Duet . . . . .

Longitudinal Fission Rate Digtribution . . .

PWAR-1 Elevated Temperature Critical Assembly Fission

Rate Distribution e s+ s e e & e s

Radial Fission Rate Distribution, Tubes A, B, Dy - D5

86

in

Page No.

10

11

14
17
19
20

22
26

28
31

36

Lo

Ll
Figure No.
21. Radial Fission Rate Distribution, Tube D; . .
22, Radial Neutron FIuX « . « « .« < . . .
23.lSketch of Crifica.l Assenmbly Showing Dimensions
24, Flow Diagram of Inert Gas System . . . . .
25. Electrical Wiring Disgram . . . . . . .
26. Heater Locations . . . . . « « .« . .
27. 'I'hemocouple i.ocations e e e o e« s e »
28. Control Rod e & o s e« o s & e e
29. Sump Tank Pipe Cap Failure . o & e a2

 

Page No.

| o
52
54
58
60
65
66
75
7

 
 

Appendix D

LIST OF TABLES

Tabie No.

l.

1z2.

13.

1k,

15,
16.

17.

18.

195.

20,

Fuel Constituents . . . .

Incremental Core Volumes

Distribution of Holes in the Reflector and Moderator

Dimensions for Reactor Assembly Sketch .

Reactivity Value of the Control Rod at Various Positions

Effect on Reactivity of Imserting the Blo Sleeves into

the End Duct Beryllium .

Control Rod Position as a Function of Uranium Concentration

and B1O Sleeve Position

Specific Mass Reactivity Coefficients

Varistion of Critical Control Rod Position and Rod Value

with Temperature « o .

Effect of Fuel Height on Rod Sensitivity
Effect of Fuel Level on Reactivity
Fast-Neutron Leakage Measurements

Conditions of Reactor During Longitudinal Fission Rate
Traverses in the Upper End Duct

Longitudinal Fission Rate Distribution in the Upper End

mct . - - - ® . - °

Positions of Radial Detector Tubes

Radial Fission Rate Distribution .

Longltudlnal Positions of Gold Foils in Cold and Hot

Agsemblies . . .+ . .« .

Radial Gold Neutron Flux Distribution in the Beryllium

Reflector o o e e e »

Dimensions of Inner Core'Sheli

Dimensions of OCuter Core Shell

88

-

-

 

16
23

25

2T
30

32
33
35
3T

59

L1
b3
b6

50

51
70
T1
Teble
21.

22.
23,
2k.

25.

89

List of Tables (cont.)

No.
Composition of B0 Pieces . . . .

Camposition of Control Rod Cermet .

Composition of Beryllium Oxide Foil Holders .

Composition and Weights of Beryllium Reflector and

Island Parts . . . e o + e

Coamposition and Weights of Core Shells

.

Page No.
83

83
83

85

 
-91-~ e
ORNL-2536 . 7%
Reactors-Special Features
of Aircraft Reactors

M-3679 (21st ed.)

 

INTERNAL DISTRIBUTION

. 1. C. E. Center 38. D. S. Billington
2. Biology Library 39. H. E. Seagren
3. Health Physics Library 40. A. J. Miller
4-5. Central Research Library 41. M. J. Skinner
6. Reactor Experimental 42. R. R. Dickison
Engineering Library 43. J. A. Harvey
7-11. Laboratory Records Department 44. A. Simon
12. Laboratory Records, ORNL R.C. 45. F. C. Maienschein
13. A. M. Weinberg 46. C. E. Clifford
14. L. B. Emlet (K-25) 47. A. D. Callihan
15. J. P. Murray (Y-12) 48. R. R. Coveyou
16. J. A. Swartout 49, W. Zobel
17. E. H. Taylor 50. F. L. Keller
18. E. D. Shipley 51. C. D. Zerby
19. A. H. Snell 52. D. K. Trubey
20. E. P. Blizard 53. S. K. Penny
21. M. L. Nelson 54. R. W. Peelle
22. W. H. Jordan 35. W. K. Ergen
23. G. E. Boyd 56. A. P. Fraas
24, R. A. Charpie 57. W. R. Grimes
25. 8. C. Lind 58. W. D. Manly
26. F. L. Culler 59. A. M. Perry
27. A. Hollaender 60. H. W. Savage
. 28. J. H. Frye, Jr. 61. W. C. Tunnell
29. M. T. Kelley 62. F. L. Friedman (consultant)
30. J. L. Fowler 63. H. Goldstein (consultant)
. 31. R. S. Livingston 64. H. Hurwitz (consultant)
32. K. Z. Morgan 65. L. W. Nordheim (consultant)
33. T. A. Lincoln 66. R. R. Wilson (consultant)
34, A. S. Householder 67. ORNL ~ Y-12 Technical Library,
35. C. P. Keim Document Reference Section
36. C. S. Harrill
37. C. E. Winters

EXTERNAL DISTRIBUTION

68-71. Air Force Ballistic Missile Division
72-73. AFPR, Boeing, Seattle
74. AFPR, Boeing, Wichita
/5. AFPR, Curtiss-Wright, Clifton
~76. AFPR, Douglas, Long Beach
77-79. AFPR, Douglas, Santa Monica
80. AFPR, Lockheed, Burbank
81-82. AFPR, Lockheed, Marietta _
83. AFPR, North American, Canoga Park
84. AFPR, North American, Downey
85-86. Air Force Special Weapons Center
87. Air Materiel Command
. .88% Air Research and Development Command (RDTAEF)
89. Air Research and Development Command (RDZPSP) -

 

 

 
 

90.
91-93.
94,

95.

96.

97.

98.

99.

100 -105.
106.
107.
108-109.
110.
111.
112.
113.
114.
115.
116.
117-118.
119.
120.
121.
122-127.
128.
129.
130.
131.
132-133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143,
144,

145

159.
160.
161.
162.
-163.

i
o et s e
Ty

AT

e T Ty

-92-

Air Technical Intelligence Center

ANP Project Office, Convair, Fort Worth

Albuquerque Operations Office

Argonne National Laboratory

Armed Forces Special Weapons Project, Sandia

Armed Forces Special Weapons Project, Washington

Army Ballistic Missile Agency

Assistant Secretary of the Air Force, R&D

Atomic Energy Commission, Washington

Atomics Intermational

Battelle Memorial Institute

Bettis Plant (WAPD)

Bureau of Aeronautics

Brueau of Aeronautics General Representative

BAR, Aerojet-General, Azusa

BAR, Convair, San Diego

BAR, Grumman Aircraft, Bethpage

BAR, Martin, Baltimore

Bureau of Yards and Docks

Chicago Operations Office

Chicago Patent Group

Curtiss-Wright Corporation

Engineer Research and Development Laboratories

General Electric Company (ANPD) (copy to M. C. Leverett
General Nuclear Engineering Corporation

Hartfocrd Area Cffice

Knolls Atomic Power Laboratory

Lockland Area Office

Los Alamos Scientific Laboratory (copy to H. C. Paxton)
Marquardt Aircraft Company

Martin Company

National Advisory Committee for Aeronautics, Cleveland
National Advisory Committee for Aeronautics, Washington
National Bureau of Standards .
Naval Air Development Center

Naval Air Material Center

Naval Air Turbine Test Station

Naval Research Laboratory

New York Operations Office

Nuclear Metals, Inc.

. Office of Naval Research
146.
147.
148,

149-158.

Office of the Chief of Naval Operations (OP-361)
Patent Branch, Washington
Phillips Petroleum Company

e
X h,'i.:':, voa

and J. W. Morfitt)

Pratt and Whitney Aircraft Division (copy to G. W. Alwang, A. Carson, W. J.

Fader, W. G. Kennedy, E. V. Sandin, and R. I. Strough)
Redstone Arsenal

San Francisco Operations Office

Sandia Corporation

School of Aviation Medicine

Sylvania-Corning Nuclear Corporation

 
164.
165.
166.
167.
168-169,
170-188.
189-213.
214,

-93 -

Technical Research Group

USAF Headquarters

USAF Project RAND

U.S. Naval Radiological Defense Laboratory

University of California Radiation Laboratory, Livermore
Wright Air Development Center (copy to R. E. Malenfant)
Technical Information Service Extension, Oak Ridge
Division of Research and Development, AEC, ORO