FFR_chap07.txt

CHAPTER 7

DESIGN AND CONSTRUCTION OF EXPERIMENTAL
HOMOGENEOUS REACTORS*

7—1. INTRODUCTION

7-1.1 Need for reactor construction experience. The power reactor de-
velopment program in the United States is characterized by the construe-
tion of a series of experimental reactors which, it is hoped, will lead for each
reactor type to an economical full-scale power plant. Outstanding examples
of this approach are afforded by the pressurized water reactor and boiling
water reactor systems. The development of pressurized water reactors
started with the Materials Testing Reactor, followed in turn by the Sub-
marine Thermal Reactor (Mark 1), the Nautilug Reactor (Mark II), and
the Army Package Power Reactor. KExperience obtained from the construce-
tion of these reactors was applied to the full-scale plants built by the West-
inghouse Electric Company (Shippingport and Yankee Atomic Electrie
Plants) and Babeock & Wilcox Company (Consolidated Iidison Plant).

Although many have argued that the shortest route to economic power
will be achieved by eliminating the intermediate-seale plants, most experts
believe that eliminating these plants would be more costly i the long run.
To quote from a speech by Dr. A. M. Weinberg [1], while discussing large-
seale reactor projects: “The reactor experiment—a relatively small-scale
reactor embodying some, but not all, the essential features of a full-scale
reactor—has become an accepted developmental device for reactor tech-
nology.”’

An alternative to the actual construction of experimental nuclear reactors
has been proposed which consists of the development of reactor systems
and components in nonnuclear engineering test facilities, zero-power critical
experiments, and the testing of fuel elements and coolants in in-pile loops.
This approach, although used successfully in the development of various
solid-fuel coolant systems, is not completely applicable to circulating-fuel
reactors because of the difficulty of simulating actual reactor operating
conditions in such experiments. In in-pile loops, for example, the ratio of
the volume of the piping system to the volume of the reacting zone is never
quite the same as in a reactor, making it impossible to duplicate simul-
taneously the conditions of fuel concentration, enrichment, and power
density. In cases where these variables are important, the m-pile loops

*Prepared by J. A. Lane, with contributions from 8. E. Beall, 5. I. Kaplan, Oak
Ridge National Laboratory, and D. B. Hall, Los Alamos Scientific Laboratory.
340
7-2] WATER BOILERS 341

an at best provide information of an exploratory nature which must be
verified In an operating fluid-fuel reactor.

A second aspect of circulating-fuel reactors, which precludes relying
solely on engineering tests and in-pile loops, is the close interrelation of the
nuclear behavior and the operational characteristics of the fuel circulation
system, which can be determined only through construetion and operation
of a reactor. Other aspects of reactor design that can be best determined
nm an operating homogeneous reactor are continuous removal of fission
products produced in the nuclear reaction and remote decontamination
and maintenance of reactor equipment and piping.

7-1.2 Sequence of experimental reactors. It is obvious from the fore-
going that the construction of a sequence of experimental reactors has been
an important factor in the development of homogeneous reactors. In this
sequence, which started with nonpower rescarch reactors, seven such
reactors have been built (not mecluding duplicates of the water boilers).
These are the Low Power Water Boiler (LOPQO), the High Power Water
Boller (HYPO), the Super Power Water Boller (SUPQ), the Homogeneous
teactor Iixperiment (ITRIE-1), the Homogencous Reactor Test (HRE-2),
and the Los Alamos Power Reactor Experiments (LAPRE-1 and -2).
[ the sections of this chapter which follow, these reactors are deseribed
m detail, and their design, construction, and operating characteristics are
compared. Their construction covers the regime of homogencous reactor
technology involving the feasibility of relatively small reactors fueled with
aqueotls =olutions of urantum. Since their construction and operation does
not include svstems fueled with aqueous suspensions of thorium oxide
and or uranium oxides necessary for the development of full-scale homo-
geneous breeders or converters, additional experimental reactors will un-
douhtedly he built.

7—2. Warter BoiLers*

7-2.1 Description of the LOPO, HYPO, and SUPO [2-4]. Interest in
homogencous reactors fueled with a solution of an enriched-uranium salt
was initiated at the Los Alamos Scientific Laboratory in 1943 through an
attempt to find a chain-reacting system using a minimum of enriched fuel.
The first of a sequence of such reactors, known as LOPO (for low power),
went critical at Los Alamos in May 1954 with 565 grams of U2?5 as uranyl
<ulfate. The uraniun, containing 14.5%, U35 was dissolved in approxi-
mately 13 liters of ordinary water contained in a type-347 stainless steel
sphere 11t diameter and 1/32 . in wall thickness. The sphere was sur-

*Prepared {rom reports published by Los Alamos Scientific Laboratory and
other sources as noted.
342 IXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cuHap. 7

rounded by beryllium oxide as reflector in order to minimize the eritical
mass of the U235 The lack of u shield and cooling system limited the heat
power level of LOPO to 50 milliwatts. A cross-sectional drawing of the
LOPO is shown m Iig. 7-1.

Iollowing successful low-power operation of the LOPO, the reactor was
provided with a thicker sphere (1/16 in.), integral cooling coils, and «a
shield to permit operation at 6 kw. Also, part of the beryllium oxide
refiector was replaced by o graphite thermal column, and holes through
the shield and reflector were provided for experiments. The eritical mass
of the modified reactor was 808 grams of U235 as uranyl nitrate at 14.09%
enrichment, contained in 13.65 liters of solution. The change from uranyl
sulfate to nitrate was made because an extraction method for the removal
of fission products was known only for the latter solution at that time. The
modified reactor, called HYPO (high power), went critical i December
1944 and operated at a normal power of 5.5 kw, producing an average
thermal-neutron flux of 10! neutrons/(em?)(sec).  The temperature of
the solution during operation reached 175°F with cooling water (50 gal /hr)
at 46°1.

Since higher neutron fluxes were desired, as well as more rescarch facilities
than available from HYPO, the reactor was further modified and renamed
SUPO (super power water boiler).

The modifications were made in two parts. The first phase, begun 1n
April 1949 and completed in ebruary 1950, improved the experimental
facilities and increased the neutron flux. The second phase, begun in
October 1950 and completed in March 1951, increased the thermal neutron
irradiation facilities, improved the reactor operation, and removed the
explosive hazard in the exhaust gases.

The first group of alterations consisted of the following:

(1) The space around the reactor was increased by enlarging the building
so that experiments could be carried out on all four sides instead of only two.

(2) The construction of a second thermal column was made possible by
eliminating a removable portion of the reactor shield. This made available
a neutron beam and irradiation facilities on a previously unused face of the
reactor.

(3) The entire spherical core assembly was replaced as follows:

(a) Three 20-ft-long, 1/4-in.-OD, 0.035-in.-wall-thickness stainless
steel tubes replaced the former single cooling coil. This inereased the
operating power level from 5.5 kw to a maximum of 45 kw.

(b) A new removable level indicator and exit gas unit was mstalled
in the sphere stack tube. The stack tube itself was made more accessible
for future modifications.

(¢) External joints were not welded, but unions of flare fittings
were used to simplify the removal of the sphere or permit pipe replacements.
T 2] WATER BOILERS 343

  
  
   

Overflow
Cadmium
Safety Curtain Level

Electrode

Safety Cd Control Rod

Electrode

 

 

Upper Pipe

 

Stainless Steel Sphere
Containing Enriched
Uranyl Sulfate (UO 950 4)

 

Be O Reflector

 

Graophite Reflector

 

 

 

 

Level Electrode

 

 

Air Pipe

-\D

Basin
L ump

 

 

 

 

 

 

 

 

 

Fig. 7-1. Cross section of LOPO, the first aqueous solution reactor.

(i An additional experimental hole was run completely through the
reccctor tangent to the sphere.  This 1is-in-ID tube supplemented the
I-in-1D “glory hole” running through the sphere.

4 The beryllium portion of the reflector was replaced by graphite.
The wll-graphite reflector gave a more rapid and complete shutdown of the
reactor and eliminated the variable starting source produced by the (y,n)
reaction on beryllium. A 200-millicurie RaBe source placed in the reflector
was used as a startup neutron source,

(5) Two additional vertical control rods were added which moved into
the sphere in re-entrant thimbles. These consisted of about 120 grams
of sintered B m the form of 9,/16-in. rods about 18 in. long. These
rods gave the additional control required by the change to an all-graphite
reflector.  Previously observed shadow effects were eliminated by the m-
ternal position of the rods and by the location of the control chambers
mnder the reactor.

i1 The reactor solution was changed from 159, U235-enriched uranyl
nitrate to one of 88.79; enrichment. This made possible the continued use
of a low uranium concentration m the solution with the poorer all-graphite
retlector. The gas evolution produced by nitrie acid decomposition was
creathy reduced, due to the lower total nitrogen content.
344 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

(7) The entire inmer reactor shicld was improved to permit higher power
operation with o low neutron leakage and also to inerease the neutron-to-
gamma-ray ntensity in the thermal columns. Cadmium wus replaced by
B4C paratiin and additional steel shielding was added.

After operating the reactor with the above modifications for about 10,000
kwh at a power of 30 kw, the following (second group) alterations were
made:

(1) The original south thermal column was completely rebuilt
improved shielding to provide many more irradiation facilities,

(2) A recombination system was construeted to handle the off-gases from
the reactor. The use of a closed circulating gas system with o catalyst
chamber of platinized alumina removed any explosive hazard in the ex-
haust gases due to the presence of hydrogen and oxygen. The operating
characteristics of the reactor were greatly improved by returning dircetly
back to the reactor as water all but a very small fraction of the gases
produced.

(3) A shielded solution-handling system was constructed to simplify
the procedure of routine solution analysis and for the removal or change of
the entire reactor solution.

The average neutron flux in the SUPQO during operation at 45 kw is
about 1.1 X 10" neutrons/(cm?)(sec), and the peak thermal flux (in the
“glory hole”) 18 1.7 X 10'2 neutrons/(em?)(sec). Fstimated values for the
maximum intermediate and fast fluxes at 45 kw are 2.8 and 1.9 X 102 neu-
trons/(em?) (see), respectively. Calculations made from fast beams emerg-
ing from the north thermal column at this same power level gave the fol-
lowing fast-flux values above 1 Mev in units of neutrons/(em?2)(sec):
(1) at sphere surface, 1.1 X 10'%; (2) at bismuth column, 7 X 10'%; and
(3) at a graphite face 1 ft in front of the bismuth column, 2 X 10%.

The production of hydrogen plus oxygen due to radintion decomposition
amounts to approximately 20 liters/min during operation of the reactor at
45 kw. These gases leave the reactor core and pass through a reflux con-
denser which removes much of the water vapor and then through a stainless
steel-wool trap for final moisture removal. A blower feeds the gas into one
of two interchangeable catalyst chambers containing platinized alumina
pellets. These chambers, operating at 370 to 470°C, recombine the hy-
drogen and oxygen, and the gas leaving the catalyst contains the water
rapor formed. A second condenser reduces the temperature of the exit
gas to that entering the catalyst chamber. A total of 100 liters/min of gas
1s eirculated continuously in the closed gas system at pressures slightly
above atmospheric, and the hydrogen concentration is kept below the
detonation limit at all points of the system. Excess pressures produced in
the gas system can be bled to the atmosphere through a 150-ft-high exhaust
stack.

with

y
7-2]

WATER BOILERS

345

The characteristics of LOPO, HYPO, SUPO, and the North Carolina
State College Water Boilers are summarized in Table 7-1.

DEsioN CHARACTERISTICS OF WATER BOILERS

TasLE 7-1

 

 

 

 

LOPO HYPO sSUPO NCSR [5]
Power level, kw 5 X 1072 5.6 45 10
Solution (111 HQO) IYOQSO:{ UOz(NO'})z IYOQ(NO3)2 UO2SO4
U235 wt., grams 765 870 870 848
- Solution volume, 13 13.65 13.65 15
hiters
- Enrichment, % 14.6 14.0 88.7 90
AMaximum thermal- | 3 x 109 2.8 X 101 1.7 x 1012 5 x 101
neutron flux
Reflector material BeO Be and Graphite Graphite
graphite
Coolant flow rate None 50 180 240
gal hr
Solution tempera- 39 85 85 &0
ture, °C
Experimental None 1 thermal 2 thermal 1 thermal
facilities column columns {'glory column
hole” and 12 exposure
tangential hole) ports

 

 

 

 

 

 

 

7-2.2 Kinetic experiments in water boilers. In August 1953, experi-
ments were performed on the SUPO by a group of scientists from the Oak
Ridge National Laboratory and Los Alamos [6,7] to determine the degree
to which a boiling (and nonboiling) homogeneous reactor automatically
compensates for suddenly imposed supereritical conditions. Previous
boiling experiments in 1951, unreported in the open literature [8], had
indicated the stability of SUPO under steady-state boiling conditions;
346 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [CHAP. 7

however, there remained considerable doubt as to the adaptability of a
reactor of this type to a sudden introduction of excess reactivity such as
might occur with a sudden inercase in pressure above the reactor. The
tests were performed by suddenly ejecting a neutron poison, consisting of
an aluminum rod containing boron carbide at its tip, and simultaneously
meaguring the neutron flux level with high-speed recorders connected to
a boron-coated ionization chamber located in the graphite reflector. The
amount of reactivity introduced was determined by the position of a
calibrated control rod. Although the experiments were interrupted by
frequent accidental serams, caused by the unsuitability of SUPO to boil-
ing at high solution levels in the sphere, the results indicated that both
boiling and nonboiling solution reactors are capable of absorbing reactiv-
ity inereases of at least 0.495 kg added in about 0.1 sec. In both boiling
and nonboiling eases, the reactor power was self-regulating, but excur-
siong were terminated more rapidly under boiling conditions. The average
lifetime of prompt neutrons in the reactor was caleulated from the initial
prompt rise in the neutron flux and found to be about 1.7 X 107 * sec.
Following a reactivity addition, the initial rate of reactivity decrease
(0.2 sec after start) was greater than about five times the rate which could
be attributed to core-temperature rise and the associated negative tempera-
ture coeflicient (0.0249 k/°C). As the gas bubbles left the core region,
reactivity decrease due to core-temperature rise Increased I relative
importance.

More recent experiments with the Kinetiec Experiment for Water Boilers
(KEWB-1), operated by Atomics International for the U. S. Atomic
Energy Commission [Y], have verified the self-controlling features of a
solution-type reactor. It was found that automatic shutdown due to the
temperature increase and formation of gas bubbles in the reactor fuel
solution occurs under all abnormal operating conditions tested.

7-2.3 The North Carolina State College research reactor [5]. The sim-
plicity of the Water Boiler reactor has made it of interest as a laboratory
tool for experimental work with neutrons and gamma rays and also to
provide training in reactor operation, and ten such reactors were in opera-
tion or planned in the United States by the end of 1957. The first college-
owned nuclear research reactor, which started operating at 10 kw in Sep-
tember 1953 at North Carolina State College, Raleigh, North Carolina,
was of this type. It was completed after four yvears of planning, design,
and construction, at a cost of $130,000 for the reactor, plus $500,000 for
the reactor building and associated laboratory equipment. It differs from
the Los Alamos SUPO in that the fuel container is a cylinder 11 in. in
diameter and 11 in. high, rather than a sphere. Its experimental facilities
include 12 access ports and a thermal column.
p—

7-2] WATER BOILERS 347

In June 1955, the reactor was shut down because of leaks which developed
in the fuel container and permitted the radioactive fuel to contaminate the
inside of the reactor shield. After a major repair job, operation of the
reactor with a new core was resumed in March 1957 at a power level of
500 watts, and the reactor has operated successfully at that level for a year.

7-2.4 Atomics International solution-type research reactors. Several
versions of the Water Boiler are being offered commercially by various
companies. The major supplier is the Atomies International Division of
North American Aviation Company which has built, or is building, 11 such
reactors. Low-power reactors are: the l-watt Water Boiler Neutron Source
(WBNS) originally at Downey, California, which was moved to Santa
Susana and modified to operate at 2 kw; a new J-watt laboratory reactor
(L-17) for Atomics International; the 100-watt Livermore Research
Reactor at Livermore, California; and a 3-watt reactor planned for the
Danish Atomic Energy Commission at Risg, Denmark. Higher power
Water Boilers, operating at 50 kw, include the Kinetic Experiment for
Water Boilers (KEWB-1) at Santa Susana; the UCLA Medical Facility
at Los Angeles, California; and reactors for the Armour Research IFounda-
tion in Chicago, Illinois; the Japan Atomic Energy Research Institute at
Tokai, Japan; Farbwerke Hoechst A. G. at the University of F rankfurt,

        
   
 
 

" Reactor Core TR |
- %Concrem Shielding X
o )

o

Neutron Exposure Facilii;y_‘_‘

TR s

 

"Control Room - - \ _

Reactor Utili‘ty Room”

Fig. 7-2. Armour Research Foundation research reactor (courtesy of Atomics
International, a division of North American Aviation Co.)
348 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

West Germany ; the Senate of West Berlin (Institute for Nuclear Research),
Germany; and the Politecnico Enrico Fermi Nueclear Study Center at
Milan, Italy.

The first solution-type research reactor for industrial use went into
operation in June 1956. The general features of this reactor, which Atomics
International built for the Armour Research Foundation at the Illinois
Institute of Technology, Chicago, Illinois, are shown in Fig. 7-2.

The exposure facilities include one 6-in.-diameter beam tube extending
radially to within 2 ft of the core tank and 4 in. in diameter from there
to the core tank; two 4-in. and two 3-in. beam tubes extending radially to
the core tank; two 2-in. through-tubes passing tangentially to the core
tank; one 13-in.-diameter tube passing through the central region of the
core; four 4-in. vertical tubes located in the reflector; one 5 ft X 5 ft graphite
thermal column with a 12-in.-square removable section. The tube facilities
consist of steel sleeves extending through the concrete shield and aluminum
thimbles or liners which reach to the immediate vicinity of the core. Each
tube facility i1s equipped with a graphite reflector plug and a dense con-
crete-and-steel shielding plug to be installed when the facility is not in use.

The horizontal thermal column is formed by a 5-ft-square column of
graphite, in the center of which are nine removable graphite stringers. A
large volume which may be used for exposures is provided between the
end of the thermal column and the inner face of a movable concrete door.
The thermal column access ports open into this volume.

To take advantage of the 50,000 curies of gamma activity produced by
the fission-product gases circulating through the gas recombiner tank,
exposure facilities are provided which extend from the subpile room into
the exposure room and into the valve room. The facilities listed below
consist of steel sleeves and aluminum thimbles which extend through the
dense concrete walls of the exposure room.

2 gamma ports, 4 in. diameter
2 gamma ports, 8 in. diameter
1 rectangular gamma slot, 6 in. X 18 in.

In addition, two 4-in.-diameter gamma ports extend from the subpile
room into the valve room. As with the beam tubes, each port is equipped
with a plug to be installed for shielding purposes when the port is not in use.

7-3. Tae HomogeENEOUs REacTOrR ExpermMExT (HRE-1) [10-13]*

7-3.1 Introduction. In 1950 the Oak Ridge National Laboratory under-
took the task of designing, building, and operating a pilot-plant fluid-
fuel reactor, the Homogeneous Reactor Experiment (HRE-1), shown in

*Based on a paper by C. E. Winters and S. E. Beall [10].
7-3] THE HOMOGENEOUS REACTOR EXPERIMENT (HRE-1) 349

 

 

 

 

¥

 

STy
A
v
5

y

=

  

Fia. 7-3. The homogeneous reactor experiment (HRE-1).

Fig. 7-3. The purpose of this reactor was to investigate the nuclear and
chemical characteristics of a ecirculating uranium solution reactor at
temperatures and powers sufficiently high for the production of electricity
from the thermal energy released. Specifically, it was designed to operate
with a full-power heat release of 0.6 to 3.5 million Btu/hr (200 to 1000 kw
of heat), and a maximum fuel-solution temperature of 482°F, yielding,
after heat exchange, a saturated-steam pressure of about 200 psi.

During the 24-month period in which the reactor was in operation,
starting in April 1952, liquid was circulated for a total of about 4500 hr.
The reactor was critical a total of 1950 hr and operated above 100 kw for 720
hr. The maximum power level attained was 1600 kw. The reactor was shut
down in the spring of 1954 and dismantled to make room for the Homo-
geneous Reactor Test (HRE-2), having successfully demonstrated the
nuclear stability of a circulating-fuel reactor. The characteristics of the
HRE-1 are summarized in Table 7-2.

7-3.2 The reactor fuel system. The reactor core consisted of a stainless
steel sphere 18 in. in diameter, through which was circulated 100 to 120 gpm
of 93% enriched uranyl sulfate dissolved in distilled water. The temperature
350 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [CHAP. 7

TABLE 7-2

CHARACTERISTICS oF HRE-1

 

Power, heat 1000 kw

Fuel U0:804 (939, enriched) in Ho0

Fuel concentration ~30 g U235 per liter (0.17 m UO2504)

U235 in core 1.5 — 2 kg

Core 18 in. diameter, stainless steel

Pressure vessel 39 in. ID, 3 in. thick, forged steel

Reflector 10 in. D20, pressurized with He

Specific power 20 kw/liter

Fuel inlet temperature 210°C

Fuel outlet temperature 250°C

System pressure 1000 psi (430 psi above vapor pressure)

Gas removal system Vortex flow through core

Radiolytic gas recombination CuSO4 (internal); flame and catalytic re-
combination (external)

Control system Reflector level, safety plates, temperature
control

Shielding 7 ft barytes conecrete

Stearn temperature 382°F

Steam pressure 200 psi

Electrical capacity 140 kw

 

 

 

 

rise of the solution passing through the core was about 72°F at a power
level of 1000 kw. The liquid was discharged from the core at a temperature
of 482°F, and cooled to 410°F by evaporating water from the shell side of a
U-~tube heat exchanger, thus generating about 3000 Ib/hr of 200 psi steam.
A canned-rotor centrifugal pump returned the fuel to the core to be re-
heated. The total volume of solution in the high-pressure system was about
90 liters, of which 50 liters were in the core. A schematic flow diagram of
HRE-1 is shown in Fig. 7-4.

A total pressure of 1000 psi was maintained in the fuel system by heating
a small volume of fuel to 545°F in a pressurizer chamber directly above the
sphere. 'The 1000 psi total pressure, which is over 400 psi greater than is
required to prevent boiling of the fuel solution, was necessary to minimize
the volume of decomposition gases.

7-3.3 The reflector system. The reflector of HRE-1 was a 10-in. layer
of heavy water surrounding the core vessel. The heavy water was pressur-
1zed with helium to within 4100 psi of the fuel pressure in order to minimize
stresses 1In the 3/16-in. wall of the spherical fuel container. Both the
reflector and the concentric core were contained in an outer pressure vessel
7-3] THE HOMOGENEOUS REACTOR EXPERIMENT (HRE-1) 351

To Stack ¢ Steom Drum
200 psi St
t o psi eum/
Gos Blower .

e

 
 
  
  

 
   
 

Rod Drive Mechuanism 120 gpm 3

   
  
  
 
   

2 Heat Exchanger
Fission Gas 1 /s . '
Adsorber Condenser Catalytic \— Fuet Cirwlating .
f\ 311y Recambiner “ 5 Pump Main Steam Valve "
Pressurizer . o
2 1000 psi 250°C
Cold T e el supnies § [ AL
(-20°C) Storage Tun 1000 psi) 7
Condensate Weighing Tanks b3 Hydrogen-Oxygen ! s Generator

Flame Recombiner
Hydrogen, Oxygen,
Staam, and Fission Gases

|
(\ ! "fl

2 D0 Cooler

 
     
  
   

   

(ondener Neutron Absorhing

Plates

  

Turbine

 
 
 

High Pressure oo .
Reducing Valve® D,0 Druin Valve

i 2
W00 psi y750¢ ) 15 psi

 
   

1 gpm
(D70 Storage Tanks
Condensote
15 osi . Return Valve
p 1000 psi (Close To 1000 psi
(. Concentrate ‘ 15 psi
{) Fuel Feed Pump Fuel) \’ DzoFeed Pump

F1g. 7-4. Schematic flow diagram, HRE-1.

of forged steel, 39 in. in inside diameter, with a 3-in.-thick wall. A 24-in.,
1500-psi standard ring-joint flange at the top of the vessel permitted
removal of the inner core.

In order to limit thermal stresses and to reduce corrosion of the steel
vessel, the reflector temperature was regulated near 350°I. About 50 kw
of heat conducted from the fuel core to the reflector liquid was removed by
cireculating the heavy water with a 30-gpm canned-rotor pump through a
reflector cooler which acted as a boiler feedwater preheater. A jet was
located in this high-pressure circulating loop, the suction of which drew a
continuous stream of gas from the vapor space above the reflector to a
catalytic recombiner so that the concentration of deuterium and oxygen
gases in this vapor space could be kept below explosive limits.

Some measure of nuclear control was obtained by changing the level of
the reflector. The level could be lowered by draining liquid through a valve
to storage tanks, or raised by starting a feed pump. This pump, which
emploved a hydraulically driven diaphragm with check valves, had a
capacity of approximately 2 gpm against 1000 psi. Its intake was connected
to supply tanks at atmospheric pressure, located below the reactor. These
tanks also served as degas chambers for the reflector liquid which was dis-
charged from the pressure vessel. Helium, water vapor, and D2 and O:
352 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHaAP. 7

gas liberated here passed upward through a condenser to a small low-
pressure catalytic bed where the ()2 and Do gases were recombined to D»O.
Cold traps operated at —20°I" were included in the reflector-system vent
lines to prevent the loss of D20 or its contamination with H20) vapor.

7-3.4 The fuel off-gas system. When the reactor was operating, with-
out copper sulfate in the fuel solution, the fuel solvent (H.O) was decom-
posed by the energy of fission to yield a stoichiometric mixture of hydrogen
and oxygen gas at a rate of 0.28 c¢fm (10 efm at STP). In addition to this
large volume of decomposition gases, there was also produced a very small
volume (20 ce/day) of intensely radioactive fission gus. If these gases had
not been removed and replaced by more liquid, excessive pressures would
soon result, and since virtually all of this gas was liberated within the core,
the displacement of fuel solution by the gas would make it impossible for
the chain reaction to continue. I'or this reason, the gas was continuously
separated from the fuel in the core by injecting the main circulating stream
tangentially near the equator of the sphere, which caused the fluid to rotate
and form a vertical cylindrical vortex approximately 1/4 1. in diameter.
The centrifugal action of the rotating fluid served to separate the decom-
position and fission gases from the liquid to the vortex, the axig of which
was aligned with the fuel outlet. A nozzle with a central opening in the
fluid outlet allowed the removal of gas from the vortex. This gas plus
about 0.8 gpm of the fuel solution was passed through the outer annulus of
a countercurrent, concentric-tube heat exchanger, which was partially
cooled by 0.8 gpm flow of fresh makeup liquid being pumped back to the
core. The cooled mixture of gas and liquid was then throttled through a
valve into a gas separator which was connected to the fuel-solution storage
tanks. The gas-steam mixture rose from the gas separator to a condenser
immediately preceding a flame recombiner, so that the gases leaving the
condenser were combustible and reunited to water in the flame of the re-
combiner shown in Fig. 7-4.

The flame recombiner is best described as an oversized Bunsen or Meeker
burner enclosed in a water-jacketed cylinder. In the HRE-I no attempt
was made to use this 40 kw of high-temperature heat, although in larger
scale reactors this energy might be used to superheat steam about 70°L.

The exit gases from the flame recombiner contained only fission products
and small amounts of unrecombined hydrogen and oxygen. They were
passed through a catalytic recombiner which contained a platinized alumina
catalyst to eliminate the traces of hydrogen and oxygen. Also, the catalytic
bed was used to react the entire gaseous output at low reactor powers when
insufficient gas was being liberated to maintain a steady flame at the burner
of the flame recombiner. The catalytic bed was followed by a condenser
and cold traps to prevent the loss of water from the system. The gas stream
7-3] THE HOMOGENEOUS REACTOR EXPERIMENT (HRE-1) 353

at this point was composed mainly of excess oxygen plus the highly active
fission gases, mainly xenon, krypton, and their decay products. The ac-
tivity of these gases was many orders of magnitude greater than the activity
which can be discharged directly to the atmosphere without the construc-
tion of a very expensive stack; therefore it was desirable to provide some
inexpensive means of storage for the dissipation of the radioactivity. This
was accomplished by passing the gases through cold traps to remove
moisture and adsorbing them onto water-cooled activated-carbon beds
which were buried underground outside the reactor building. It is estimated
that the equilibrium activity of the gases held on the carbon bed was 400,000
curies, *

The adsorption efficiency of the charcoal, even at ground temperature,
was good enough to prevent a discharge of activity greater than a few
curies per day. However, even this amount of activity had to be diluted
so that the atmospheric concentration at ground level was not greater
than 10718 curies/cc of air. Dilution was accomplished by feeding the
active gas into a 1000-cfm ventilating air stream from the reactor shield
and then to a 100-ft-high stack. During operation the gaseous activity
inside the stack barely exceeded inhalation tolerance.

7-3.5 Fuel concentration control. The condensate which was removed
from the vapor-gas mixture upstream of the recombiner was returned either
to the fuel storage tanks or to weighed holding tanks. The accumulation
of water in the holding tanks provided a means of increasing the concen-
tration of fuel in the storage tanks underneath the reactor. Since fuel was
pumped continuously from the storage tanks to the high-pressure system
by means of a duplex-diaphragm type pump at a rate of 0.8 gpm, it was
possible to vary the concentration of the fuel which circulated through
the reactor. Figure 7-5 shows how the core temperature varied with fuel
concentration, in g/kg HoO. Furthermore, since the operating temperature
of the core was controlled by the fuel concentration as shown in this figure,
the operator had a convenient means of adjusting the solution temperature
to the desired level. This feature of variable concentration was employed
during startup of the reactor when the concentration had to be changed
by large amounts, and also during steady operation for small changes in
temperature. When sudden dilution of the fuel was desired, as in the case
of a complete shutdown, the condensate holdup tanks were quickly emptied
through a drain valve into the fuel-storage tanks, or condensate was pumped
directly into the core.

The steep slope of the curve in Fig. 7-5—i.e., the large negative tempera-
ture coefficient—was a feature which was extremely important from safety

*One curie equals 3.7 X 101¢ disintegrations per second.
354 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

 

YT T T T T T T

40 |—
39—
38 }— |
37 |— ]
36— —
35—
34—
33—
32 |—
31—

30 +—

Fuel Concentration, g U235/kg H,0

  

29—

  

Theoretical - ~4— Experimental
28—

27}
26—
25—
gal

0 25 50 75 100 125 150 175 200 225 250
Temperature, °C

 

 

 

Fic. 7-5. Dependence of eritical fuel concentration on temperature in HRE-1.

and power-demand standpoints. Tor instance, a temperature rise of only
15°C was necessary to overcome a reactivity increase of 19, an amount
considered very dangerous in most solid-fuel reactors.

7-3.6 Power removal. The steam generated in the fuel heat exchanger
was fed to a conventional multistage condensing turbine generator rated
at 312 kva. With the reactor operating at 1000 kw and 250°C, a sufficient
quantity of steam at 200 psi was produced to generate about 140 kw of
electricity. Steam leaving the heat exchanger was first passed through a
time-delay drum with a radioactivity monitor at the inlet and a quick-
closing valve at the outlet to prevent the escape of activity into the turbine
system in the event of boiler tube failure. Small feed pumps returned the
condensate from the turbine to the boiler.

Upon increase in generator load, the turbine governor opened the
turbine throttle valves, increasing the steam demand, lowering the steam
7-3] THE HOMOGENEOUS REACTOR EXPERIMENT (HRE-1) 305

pressure and temperature, and reflecting itself into increased cooling of
the uranium solution, which automatically increased the reactivity of the
core and completely compensated for this increased load.

7-3.7 Internal-recombination experiments. The use of copper dissolved
in the reactor fuel for the complete recombination of radiolytic gas was
successfully demonstrated in the HRE-1 [14]. Copper ion was added as
copper sulfate on four occasions, increasing the copper concentration to
10, 25, 75, and 1509 of that necessary for complete recombination (i.e.,
6.6 g CuSOy4/liter) in a static system at 250°C, at 1000 psig total pressure,
and at a uniform power density of 20 kw/liter. The Investigation was
conducted at temperatures from 185 to 260°C, at pressures from 765 to
1200 psig, and at power levels as high as 1600 kw. In the main, the copper
behaved as expected from static bomb tests. The highest power level for
which all the gas was internally recombined was 1350 kw. In the course of
all copper experiments, including 350 hr of operation at the highest copper
concentration, 0.062 molar, no deleterious effects due to the presence of
copper were observed.

7-3.8 Nuclear safety. Although operating experience later verified early
predictions of the inherent safety of this reactor, at the time of design it
was considered judicious to incorporate conventional safety devices in the
reactor for protection against potentially dangerous situations which might
arise during low-power operation and until the dynamic stability had been
demonstrated by experiment. Safety measures in the order of their auto-
matic action as installed to limit reactor power or power doubling time
were:

(1) Two magnetically coupled safety plates, worth about 45 g of uranium,
which by falling in 0.01 sec caused the reactor temperature to be lowered
from 250°C to approximately 243°C.

(2) Dumping of the reflector.

(3) Dilution of the fuel (this was the normal shutdown procedure).

(4) Stopping of the steam extraction hy closing either a steam valve or
the turbine governor.

(5) Draining the fuel solution to the noncritical-geometry tanks below
the reactor.

IIxperiments demonstrated that the reactor was extremely fast-acting
with respect to limiting power surges and led to the belief that mechanical
control devices were unnecessary [15]. These consisted of a series of kinetic
experiments in which the power responses to reaetivity increases were
observed. Tirst, the entire range of normally available reactivity increases
—fuel concentration, rod withdrawal, and reflector level—was tested with
initial power levels as low as 10 watts and reactivity rates up to 0.05% per

 
356 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHap. 7

second. Then, in order to provide more drastic tests, the main fuel circu-
lating pump was stopped, the reactor was maintained at a low power and
at high temperature, but the heat exchanger was cooled about 100°C.
When the pump was restarted, the cold fuel from the heat exchanger was
rapidly injected into the core, producing a rate of reactivity increase of as
much as 0.8% k. per second. The results of two experiments in which
only the initial powers differed are shown in Fig. 7-6; the power increased
in a period as short as 35 msee, reaching a peak of 10 Mw in 1 see, and then
approached the equilibrium power demand within 0.2 sec after the peak.
The calculated pressure rise associated with the peak power of 10 Mw
was only 5 psi.

In these experiments the worst combination of circumstances was im-
posed on the reactor. It was successfully demonstrated that the HRE-1
was sufficiently stable to withstand nuclear transients greater than those
expected from operating errors.

7-3.9 Leak prevention. A major problem in the HRE-1 was to main-
tain absolute leaktightness in all components. The radioactivity of the
solution during operation was about 30 curies/ce. Twenty-four hours
after shutdown the activity was about 3 curies/ce. With these high
activities, the total leakage from the system had to be kept below 1 cc per
day. A much better performance than this was attained through the use
of canned-rotor pumps, double tube-sheet exchangers, and bellows-sealed
valves.

All welded joints were made with extraordinary care and tested by sev-
eral nondestructive methods before being approved for use. Flanged
joints were assembled with stainless steel ring gaskets of oval cross section

TABLE 7-3

HRE-1 ConsTrUcTION CoOST SUMMARY™

 

 

 

Total %% of total
Building $ 300,000 27.5
Fuel and reflector equipment and piping 420,000 38.2
Instrumentation 190,000 17.6
Shield 110,000 9.7
Power system 80,000 6.9
Total construction cost for HRE-1 31,100,000 100.0

 

 

 

*These costs include material, labor, and allocated overhead.
7-3] THE HOMOGENEOUS REACTOR EXPERIMENT (HRE-1) 357

 

   
 
 

 

 

 

 

 

T T T T T T
103 — —
3 107 — | ]
= |
- Exp 25 |
3 : J
o 10 J—  Start Pump ! e
s {Core Temp 177°C) |
G ! Traced From Brush
g 1 = / Recorder ]
o fl ———- Inferred From Exp 25
Start Pump /I
101 b= {Core Temp ’/ —
170°C) /
———e~___ R4
olEezetT L |
0 1 2 3 4 5 6 7 8 9

Time, sec

Fic. 7-6. Power response of HRE-1 during reactivity increase of 0.8%, k./sec.

and each joint contained a leak detection device for constant monitoring
(see Fig. 7-13). In addition, the ventilating air which flowed through each
equipment compartment was monitored constantly with gamma-radiation
detection devices. Although several leaks were experienced in the startup
phases, no leakage was found during the final 12-month period.

7-3.10 Shielding. As with all nuclear reactors, personnel had to be
protected from the high radiation levels which existed in the vicinity of
the reactor core. In the HRE-1 this protection was provided by a 7-ft-
thick shielding wall of high-density concrete. An increase in the density
of the concrete from 2.3 to 3.5 g/ce resulted from the use of barium sulfate
ore as the aggregate material. The shield was a departure from most reactor
shields in that it was constructed of loosecly stacked block with only the
outer 16-in. layer of blocks being mortared.

7-3.11 Construction cost. The total construction cost of the reactor
was $1,100,000, which did not include the cost of fuel and heavy water.
Table 7-3 is a summary showing the cost of various parts of the system.

7-3.12 Maintenance. The maintenance of homogeneous reactors is
greatly complicated by the radioactivity of the parts, and in the HRE-1
it was often necessary to decontaminate equipment and to provide tem-
porary protective shielding before repair work could be done. In most
cases the repair work had to be done with long-handled tools, which made
the job even more difficult. By the use of decontamination, shielding,
and extension tools, it was possible, however, to make a number of major
358 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHaP. 7

repairs in radiation fields as high as 2000 r/hr, without exposing personnel
beyond accepted tolerances. In this regard remote viewing devices such as
mirrors, binoculars, and a Polaroid Land camera were found to be invalu-
able tools. The main circulating pump was replaced or repaired three
times under such conditions and the diaphragm feed pumps twice. In fact,
in no case of a breakdown was it impossible to make the necessary repairs.

7-3.13 Dismantling the HRE-1. After final shutdown-of the TTRE-1
the reactor was decontaminated in preparation for disassembly. Over a
period of 30 days, starting with activity levels of the order of 1000 r/hr, the
activity was reduced sufficiently to permit dismantling of the system with
long-handled tools [16].

The decontamination treatment consisted of repeated washing alter-
nately with 359, HNO3 and aqueous solutions of 109, sodium hydroxide,
1.59 sodium tartrate, and 1.59; hydrogen peroxide. The over-all decon-
tamination factors were 22 to 25, including decay, but the factor for
decontamination with a single reagent was only between 1 and 2.25.
Large amounts (of the order of 10? curies) of cerium, zirconium, barium,
lanthanum, strontium, niobium, and ruthenium were removed. The
significant contaminants remaining were zirconium and niobium, which
were bound in the oxide film. Although these could have been removed
from the system by descaling the oxide corrosion film, which would have
given a further decontamination factor of approximately 100, such a
treatment would have made it impossible to determine that no significant
corrosion had taken place during nuclear operation.

7-3.14 Critique of HRE-1 [17]. After the HRE-1 was put into opera-
tion, personnel associated with the Homogencous Reactor Project were
asked to suggest ways in which the design and construction of the reactor
and the associated development program might have been improved.
More than one hundred specific design changes were recommended, many
of which related to the difficulty of operating the reactor on a continuous
basis and the need for repairing and/or replacing faulty equipment.
Among the many possible improvements recommended were the use
of high-pressure catalytic recombination; external gas separation (ie.,
nonvortex core flow); spacing of equipment for easier inspection and main-
tenance; shielded, waterproof instrument lines; instrumentation for more
accurate fuel accountability; improved feed pumps; and provision for ob-
taining meaningful corrosion data.

The general conclusion reached was that the reactor construction schedule
(16 months) was too accelerated to allow good design and construction
practices to be put fully into effect. The component development and
testing program, in particular, suffered by the short time schedule in that
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 359

pieces of equipment such as valves and pumps were not completely tested
before being used in the reactor.

The thorough analysis of the HHRIE-1 design provided a basis for the
subsequent design of the HRE-2, in which many of the suggested improve-
ments were incorporated. These included: (a) greater accessibility of
equipment, (b) provision for flooding cells, (¢) better shield construction
with metal walls to permit decontamination, (d) more accurate means for
measuring fuel inventories, and (e) elimination of screw-type fittings.

7-3.15 Summary of results. As a result of the operation of the HRE-1
and of the extensive experimental program conducted with it, several un-
certainties were resolved regarding the nuclear and chemical behavior of
aqueous homogeneous reactors at the high temperatures and high pres-
sures required for power generation. Included were demonstrations of (1)
a remarkable degree of inherent nuclear stability, a result of the very large
negative temperature coefficient of reactivity, (2) the elimination of the
need for mechanical control rods as a consequence of this inherent stability,
(3) flexibility and simplicity of fuel handling, (4) stability of the fuel,
(5) the ability to attain and maintain leaktightness in a small high-pressure
reactor system, (6) the safe handling of the hydrogen and oxygen produced
by radiation decomposition of the water, and (7) the direct dependence of
reactor power upon turbine demand.

7-4. Tur HomocENEOUS REACTOR TEST (HRE-2)*

7—4.1 Objectives. The objectives of the Homogeneous Reactor Test
(HRT-2) are (1) to demonstrate that a homogeneous reactor of moderate
size can be operated with the continuity required of a power plant, (2) to
establish the reliability of engineering materials and components of a size
which can be adapted to full-scale power plants, (3) to evaluate equipment
modifications which will lead to simplifications and economy, (4) to test
simplified maintenance procedures and in particular underwater mainte-
nance, and (5) to develop and test methods for the continuous removal of
fission and corrosion contaminants.

7-4.2 Reactor specifications and description [15,18]. The bases for
the design of the reactor, which are summarized in Table 7-4, were selected
early in 1954 and were intended to take fullest advantage of the progress
in chemistry, materials and component development, and the experience

*Based on information supplied by S. Il. Beall and 8. I. Kaplan and reports by
members of the Oak Ridge National Laboratory as noted.
360 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHaPp. 7
with the HRE-1. In order that the objective of a significant test of the
engineering feasibility of a large power station be satisfied, it was necessary
that the physical size of the reactor and its auxiliaries be increased appre-
ciably beyond that of HRIS~1. To hold the cost within reasonable bounds
for an experiment, it was decided to limit the power output and thus the
expense for heat-removal equipment, and also to install the reactor in the
building which had previously housed HRE-1, permitting the use of many
of the existing site facilities. The size of the reactor core represents a
compromise between two objectives, attainment of high specific power
required for economy in a large plant and evaluation of the fabrication
and durability of a zirconium-alloy core tank. The power output was,
therefore, set at 5000 kw (heat) with the possibility of a maximum
10,000 kw, and the core diameter at 32 in. Although these together result
in a low specific power of 17 kw/liter in the core at 5000 kw, this was con-
sidered acceptable, since operability at a relatively high specific power of
30 kw/liter had been demonstrated in the HRE-1. Another factor affecting
the selection of core diameter was the opinion of fabricators that current
technology would be exceeded for a zirconium vessel larger than 32 in.

TaABLE 7-4

DEesicN Bases ror HRE-2

 

 

Power, heat
Temperature, core outlet
Pressure

Core diameter

Core solution

Blanket

Fuel circulation rate

Blanket circulation rate

Core flow pattern

Core construction material
System construction material
Radiolytic-gas removal
Radiolytic-gas recombination

Fission-product-gas disposal
Control

Normal

Safety

 

5000 kw

300°C

750 psi in excess of vapor pressure (see text)

2000 ps1 maximum total pressure

32 In.

U02804 1n DO ("'10 g U235 per hter)

D20

400 gpm

230 gpm

Straight-through

Zircaloy-2

Type-347 stanless steel

External pipeline separator

Low-pressure system: platinized alumina
cateflyst

High-pressure system: CuSO4 in solution

Decay on activated carbon

Variable solution concentration
Temperature coefficient

 

 
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 361

The increase in fuel temperature from 250°C in the HRE-1 to 300°C was
based upon the more favorable corrosion resistance of both stainless steel
and Zircaloy-2 to dilute uranyl sulfate at the higher temperature and the
possibility of improved thermal efficiencies. Also the temperature at which
the two-liquid phase region appears is higher for the more dilute fuel, per-
mitting this increase.

TABLE 7-5

HRE-2 DEsicN PARAMETERS

 

 

 

Core Blanket

Power, heat, kw 5000 220
Pressure, psi 2000 2000
Vessel

Inside diameter, in. 32 60

Thickness, in. 5/16 4.4

Material Zircaloy-2 Stainless-steel-clad

carbon steel

Volume, liters 290 1550
Specific power, kw/liter 17 0.14
Solution UOQSO4~D20 Dzo
Uranium concentration, g of

U235 per kg D20 9.6 0
Circulation rate, gpm 400 230
Inlet temperature 256°C 278°C
QOutlet temperature 300°C 282°C
Volume of gas generated,

ft3/sec at STP 0.96 0.013

ft3/sec at 2000 psi, 280°C 0.015 e

 

 

 

 

 

The design pressure of 2000 psi resulted from the necessity for an over-
pressure on the system to prevent boiling, to reduce the volume of gas in
the core, and to increase the efficiency of the copper catalyst. A maximum
pressure of 750 psi was thus provided in excess of the vapor pressure of
water of about 1250 psi at 300°C.

The thickness of the blanket of 14 in. between the core and pressure
vessel was selected as a compromise between neutron leakage and the use
of a simple pressure vessel of dimensions within the means of standard
fabrication techniques.
362 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

The fuel circulation rate of 400 gpm was based upon heat-removal
requirements and the availability of suitable pumps. Pumps capable of
this output had been successfully operated under comparable condi-
tions.

The vortex-type flow through the core of HRE-1 was abandoned in
favor of straight-through flow because hydrodynamic experiments demon-
strated the pressure drop of the former to be excessive for larger cores.
As a result, the extraction of radiolytic gas directly from the core was
excluded, and an external gas separator in the exit pipe from the core was
used to separate gases produced in the core. The amount of these gases
depends on the operating temperature and pressure and the CuSO4 con-
centration. The excess gases are recombined at low pressure but by means
of a platinized alumina catalyst instead of combustion in a flame-type
recombiner as in the HRE-1.

In stagnant fuel lines external to the core, the radiolytically gencrated
oxygen is insufficient to replace the oxygen consumed in reacting with the
stainless steel to form metallic oxides. Since an oxygen deficiency induces
hydrolytic precipitation of the uranium, approximately 2 liters/min (STDP)
of gaseous oxygen are injected into the fuel feed stream, maintaining a
concentration of approximately 500 ppm in the high-pressure system.
The disposal of the fission-product gases which are stripped from the fuel
by the excess oxygen and radiolytic gas is accomplished by adsorption
and subsequent decay on beds of activated carbon. Experience with this
method of disposal was completely satisfactory in the HRE-1.

To reduce the xenon content of the core solution and minimize the
catalyst-poisoning effect of fission-product iodine on the rccombiners, an
iodine-absorption bed of silver-coated wire mesh was installed in the off-
gas line between the fuel dump tank and the recombiners. The bed removes
over 999, of the lodine passing through it, and will retain the absorbed
iodine at temperatures up to 450°C.

Type—347 stainless steel was designated as the material of construction
for all of the reactor except the Zircaloy-2 core vessel. Titanium was
chosen to reinforce certain points of high turbulence, such as the pump
impellers and the gas separator. Previous corrosion and welding experience
with this grade of stainless steel, also used for HRE~1, has been excellent.

Based on HRE-1 experience, it was decided to eliminate mechanical
control devices and depend entirely on varying the fuel concentration for
shim control, on temperature coefficient for transient nuclear changes, and
on dumping the fuel solution for rapid shutdown when required.

The design parameters of HRE-2 are summarized in Table 7-5.

The flow diagram for the reactor is illustrated in Fig. 7-7. Since the fuel
and blanket systems are virtually identical—the only significant differences
being the absence of a blanket iodine separator and the larger vessels
THE HOMOGENEQUS REACTOR TEST (HRE-2) 363

7-4]

wajysAg
2INSSald MOT Emmcwmmv
jajup|g ©

R el mc:c_:m___u

Z-AUH J0 199YsMO "L/ 'OId
SAIDA

18ysupu]) -, O

&

w Jojpiodoay
syun] sboiog ll

R

    
          

 

10jp10dBA]

    

sup] dwng

               
 

413|007y

    

SAIDA dwing o’

Ngr,  19]00) P> 431003 | 722,
~

aalpA dwng

          

 

            

   

A—. <
SEERN

Joppindag
juswuinyuy

        
       

y

 

abBing i "5 , . =223 0 [auiquioaay

&

 

SR *

T ~18zZ1Inssaly

           

 

                 

         

 

     

m . 2 10)DIaudn) |
< \~yn . - P 234G g, 2BURYIXT tDOH mmc.._flw '
EN::@&E§~M g l&* umog-e] ¥4 2josuspuo’) m
B g T .,
- : v u .‘«s ] 1asuapuon) i
: 7 o \¥

~

 

   

- siaqiospy
npoud uoissiy

 
  

< umoq-§91 pag uoudiosqy 1
aBung e BUIPO| PUD JaUIqUWOI3Y W

     

  

  

= aiosuspued  F7777) _
T T~ e uonjog yerjupig pue 2103 I waysAg \vf\\\\\ A wnnapa
T osuigun) o wosyg soay- 1D FETNE
pinbi puo sno P4 DHO i¥up|g
o0 [ yapig o) ¢
364 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

necessary to accommodate the greater volume of blanket fluid—the entire
blanket system is not shown. The fuel system is described as follows,

In the high-pressure system the fuel solution is pumped into the bottom
of the reactor core at 256°C and is heated to 300°C as it proceeds upward
to the outlet pipe. At the top of the outlet pipe is attached the pressurizer
in which condensate is electrically heated to a maximum temperature of
335°C to produce 2000-psi steam. The fuel flows past the pressurizer to
the gas separator, where directional vanes cause the fluid to rotate suffi-
ciently to separate the radiolytic gas (D2 and O2), the excess O2, and fission-
product gas. The separated gas forms a vortex along the axis of the pipe
and is bled to the low-pressure system. The reactor solution continues
from the gas separator to the U-tube primary heat exchanger, where it is
cooled from 300°C to 256°C by transferring heat to the boiler feedwater
surrounding the tube bundle. The 244°C, 520-psi steam produced on the
shell side of the heat exchanger is bled partially to the small (345-kva
output} turbine generator remaining from the HRE-1 and partially to an
air-cooled steam condenser. The uranyl sulfate solution flows next to the
intake of the 400-gpm canned-motor circulating pump and thence is pumped
to the core for reheating. The blanket fluid follows an identical cycle at a
flow rate of 230 gpm. This lower flow rate was based on a pump of the same
horsepower but designed to circulate a thorium oxide suspension, rather
than pure Ds0.

The gases and some entrained liquid removed by the gas separator are
transferred to the low-pressure system through a “'letdown heat exchanger,”
a jacketed pipe which cools the gas-liquid mixture to 90°C. A valve
downstream of the heat exchanger throttles the gas-liquid stream to
atmospheric pressure. The mixture then discharges into the “dump”
tanks, which have sufficient capacity to hold all the reactor liquid. An
evaporator built into the dump tanks provides continuous mixing and,
more important, steam for dilution of the deuterium and oxygen below
the explosive limits. The gas-and-steam mixture flows upward through the
iodine bed to the catalytic recombiner, in which the deuterium and oxygen
react on a bed of platinized alumina pellets to form water vapor. The
heavy water is condensed by the shell-and-tube condenser following the
recombiner and normally flows back to the dump tanks. However, the
water may be diverted to weighed storage tanks in case it is desired to
change the concentration of the fuel solution. Water which is returned
to the dump tanks is mixed with the excess fuel solution stored there (ap-
proximately 25 gal) and then fed to the intake of a sealed-diaphragm
injection pump, which returns the liquid to the high-pressure circulating
system at a rate of about 1 gpm, thus constantly replacing the liquid
removed via the gas separator.

The small volume of intensely radioactive fission gas plus the excess
THE HOMOGENEOUS REACTOR TEST (HRE-2) 365

7-4]

“Jourejuod JodsA pus ppaIys Z-JHH '8-L ‘PIg

        

             

 

         

     

 

-
S Lo s (70 “ut ) siog Budicjuiay oIg "ul-t Buippo) 9245 ss3jutig
A Pl soo|4 Juawdinby Aipyjixny S
ey ™ [eyems 19IDM PUD pung sajliog N o -
Al N T ) fi B ] 7y A .
A - - .J/ Iz\nt.fi..“ Y i
“ DDA L
i “, > .cED_oU adiy "ui-¢ |

poayyng 4 ” DR~

X
4 0’618 uUoHDAR|] O}
0.—0._UCOU >._UC_mu._o

 

R .

 

U0 M ST

 

 

 

 

{a1910u07) sa4dung) 49Apq Bn|g do)

{s18A0q Bn|q usamjag) Joayg |peg
3606 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

oxygen remaining after condensation of the re-formed heavy water is
dried in cold traps at —23°C and sent to the beds of activated carbon for a
period of decay. Gas leaving the bed is diluted with 1400 ¢fm of air and
discharged to the atmosphere from a 100-ft stack.

Samplers are provided to sccure small quantities (5 ml) of the fuel and
blanket liquids from the high- and low-pressure systems for chemical
analysis. These units are located in bypass lines; material circulated
through them is trapped by closing the sampler inlet and exit, after which
the contents are discharged into a portable container through a drain
valve.

All the primary reactor equipment is located in an underground, box-
like, steel tank called the “shicld pit,” shown in Fig. 7-8. The design of the
shield pit was influenced by several factors, including a requirement for
accessibility and flexibility because of the experimental nature of the in-
stallation, provision for complete containment of the contents of the re-
actor should a leak develop or should the pressure vessel or heat exchangers
rupture, efficient utilization of the space within an existing structure, and
capability of flooding with water for maintenance or replacement operations.

The reactor shicld pit occupies the center high-bay area of the building
and is constructed of 3/4-in. welded steel plate reinforced in such & manner
that an internal pressure of 30 psi will be contained. This pressure corre-
sponds to the instantaneous adiabatic release of the entire contents of the
reactor svstem. The chemical processing cells, each 12 ft wide by 25 {t
long, are designed similarly.

The upper surface of the blocks forming the roof of the shield pit 1s at
ground level. This roof is made of high-density concrete 5 ft in total thick-
ness and consists of two layers of removable slabs with a completely welded
steel sheet sandwiched between the layers and extending across the top of the
pit to form a gastight lid. The roof blocks are anchored to the girders and
supporting columns by means of a slot-and-key arrangement, shown in
Fig. 7-8. The vertical columns are embedded in a concrete pad which 1s
3 ft thick and is heavily reinforced with steel.

The wall hetween the reactor pit and the control area is a hollow box
5% ft wide, constructed of 1/2-in. steel plate welded to the north side of
the reactor shield tank. It is filled with high-density barytes, sand, and
water. The use of the fluid shield between the reactor and control-room
areas allows flexibility in the locations of service piping and instrument or
electrical conduits. All lines leaving the reactor tank are welded mnto the
shield wall; conduits are connected iuto junction boxes inside the pit with
gastight seals on the individual wires.

The design is such that nowhere outside the shield will the radiation
dosage exceed 10 mrep/hir when the reactor is at 10 Mw. l'or the purpose
of decreasing the neutron activation of equipment inside the pit, the re-
THE HOMOGENEOUS REACTOR TEST (HRE-2) 367

7-4]

lasuspuor)
wnaig

0
G
3
o
o
3
w
®
2

JaUiquIocoay jajun|g

‘matA ur[d ‘QuowedurLie juauodwiod - "6-2 PLI

yue] dwng
43N0 J9up|g

(1oxuoig)
19Bupy>x3 josy

7o ML Auo
dwngq jaxjuo|g

°

sdosj p|od
2un|g
101003y
Juoy
8ipsUapuUO jaxyuD|g

SSAIDA

sdwing
pP334 i93ub|g

——

J9suspuor
xn|45y

43Z1Inssaly

SBAIDA

sdwny
pead |20y

juo] dwng .

JanQ |any

sdoa) pjoy jeng

{]ony)

JeBubyox g joay

lasuspuc)
Jaulquodey |any

 

 

 
368 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

F1a. 7-10. HRE-2 container looking southeast {at 509, completion).

 

 

F1a. 7-11. Reactor tank with shielding plugs in position during hydrostatic test.
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 369

actor vessel 15 surrounded by a thermal neutron shield consisting of a steel
tank with a 2-ft-wide annulus filled with boron ore and water.

The arrangement of reactor components within the main reactor pit is
shown in Fig. 7-9. The reactor is near the center of the pit, enclosed by
the 2-ft-thick thermal shield. To the left of the reactor is all the fuel-
system equipment, and to the right is the blanket equipment.

Most of the high-pressure components are located close to the pressure
vessel. The pressurizers are positioned directly above the reactor; the gas
sepuarators are in the S-shaped outlet lines to the steam generators on the
south side of the pit. The areas to the left and right of the reactor in the
center bay are reserved for future equipment modifications or additions.

To the far left is grouped all of the fuel low-pressure equipment, including
dump tanks, recombiner, condensate tank, and cold traps. These compo-
nents are assembled on a rigid structural-steel frame; the blanket low-
pressure equipment is to the far right.

Insofar as possible, valves are situated close to the control-room wall
so that air lines and leak-detector lines can be kept short. The arrange-
ment of the valves is such that all flanges can be easily disconnected from
above.

7-4.3 Schedule of construction [19,20]. Construction of the HRE-2
was started in July 1954, immediately after the dismantling of HRE-1.
The initial step was the excavation of a large hole beneath the building
(which had previously housed the HRIZ-1) for the large rectangular steel
tank (60 ft long, 304 ft wide, and 25 ft deep) which contains the reactor and
its associated equipment. [igure 7-10 shows this tank at approximately
209 completion,

The north wall of the reactor tank, seen to the left in Fig. 7-10, is com-
mon to both the reactor tank and the control-room area, and it is through
this wall that the many service, instrument, and electrical lines which must
interconnect these two areas pass. The next step in the construction was to
install the approximately 600 lines which penetrate this 5i-ft-thick wall.

Figure 7-11 shows the tank after the complete roof structure had been
assembled and welded closed as it was prepared for a hydrostatic test. In
this test the tank was filled with water and then pressurized to give the
equivalent of a 30-1b internal pressure at all points within the tank. Strain
gauges were attached at many points so that the complicated stress pattern
could be studied in some detail for assurance that the tank was safely
within design limits.

While the reactor tank and control-room areas were being constructed,
reactor equipment was being procured and constructed at several places.
Much of the equipment required in the low-pressure system had been
inspected and tested when the container was completed. In November
[cHAP. 7

 

 

.

a2

£

<

=

g

mm ....,,. = A

s W

e
W

1
g
o g

of HRE

TAL REACTOR DESIGN AND CONSTRUCTION

MEN

EXPERI

370

 

-12. View

Fic. 7

 

Fia. 7-13. Artist’s coneept of homogeneous reactor test, HRE-2.
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 371

 

 

 

 

  

 

 

   

  

DO
| Supply
To Oxygen System
{3000 psi) Vent to Air Control Area Reactor Cell 8
Pressu.e Drop Alarm ] 2
Seal Weld -
—s i< s
SIFQOF:'LJC:;?j y femmamel < . Secl Weld Indi‘vfiual Test .
{ e e, Point Line -
Level Cnnfro\\\ ] Pressure Goge —+ 9 B
Y ! North Shield wall Flanged
T ot Joint
I
<
ot i
e~
e e Legend:
e et ——{>fme Valve Normolly Open
d -t Valve Normally Closed

 

 

 

 

 

- Sectionalizing Heoders

Fic. 7-14. HRE-2 leak detection system.

Pressurizer Vessel

1955, these parts and the thermal shield surrounding the reactor pressure
vessel were installed. Figure 7-12 shows the reactor core and pressure-
vessel assembly just before installation in January 1936.

The heat exchangers were subsequently installed and followed by the
main circulating pumps so that the high-pressure piping which connected
the pumps, heat exchangers, and pressure vessel could be attached. This
work occupied most of the months of February and March. Construection
of the reactor was completed in May 1956. ['igure 7-13 is an artist’s
concept of the completed reactor.

7-4.4 Nonnuclear testing and operation. Pretesting, operation of the
reactor as o nonnuclear facility, and a lengthy flange-replacement job
occupied the period from completion of construction in May 1956 to De-
cember 1957. A chronological summary of the events associated with the
nonnuclear operation of the reactor during this period is shown in Table 7-6.

From Table 7-6 1t can be noted that preoperational testing of HRE-2
was interrupted by stress-corrosion cracking difficulties, which were caused
by chloride ion contamination in the stainless steel tubes that are used to
detect and prevent leakage of radioactive solution from flanged joints [20].
Figure 7-14 shows how an individual leak-detector line is attached to the
groove of the ring-joint flange. The tubes from all the flanges terminate
at a valve header station in the control room. Normally this system is
kept pressurized with water to a pressure of 300 to 500 psi above the system
pressure. A leak m any flange results in leakage of water from the header
and o loss in pressure, which actuates an alarm at a fixed level above the
fuel or blanket pressure. This is normally o sensitive and satisfactory means
372

EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION

[cHAP.

TABLE 7-6

SumMaryY oF HRE-2 NonNNUCLEAR OPERATION

 

Period

Test or event

 

 

 

May 1956

June 1956—
July 1956

July 1956—
August 1956

August 1956
October 1956

October 1956-
November 1956

December 1956--

January 1957

January 1957-
March 1957

April 1957-
August 1957

August 1957-
September 1957

September 1957
October 1957

November 1957—
December 1957

 

3000-psig hydrostatic test of high-pressure system and
750-psig test of low-pressure system

Cleaning of piping systems with 3%, trisodium phos-
phate followed by 59 nitric acid and initial operation
of pumps. Tests for dump-tank entrainment and
efficiency of catalytic recombiner

Initial tests of equipment removal and underwater
maintenance

Flushing of flange leak-detector tubes to remove chlo-
ride contamination

Initial nonnuclear operation of reactor at 280°C and
2000 psig. Thermal cycling of flanged joints

Removal of typical flanges for metallographic examina-
tion to detect possible stress-corrosion cracks. Further
tests of remote-maintenance tools

Further operation of the reactor with water and with
depleted uranium at various conditions of tempera-
ture and pressure

Removal, inspection, and replacement of flanges and
leak-detection tubing

Recleaning of system with trisodium phosphate and
nitrie acid solutions and hydrostatic testing

Final operation at design conditions with water and
with depleted uranium. Final leak test of reactor
piping with radioactive tracers

Final insulation of reactor piping; installation of new
refrigeration system and new iodine absorption bed,
followed by nonnuclear operations with heavy water.
Criticality achieved December 27, 1957

 

 
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 373

of preventing the leakage of radioactive liquid from the reactor (the leak-
detector fluid leaks out instead) and detecting the leaks when they do occur
(by measuring the pressure or volume loss of the leak-detector fluid).
Volume changes in the header can be read to 41 cc from o graduated scale
next to each header sight-glass. DBy observing level changes at regular
intervals, noting which lines are isolated from the header, leaks of less than
2 ce per day can be detected.

Since this system is a secondary portion of the reactor, the leak-detector
tubing unfortunately did not receive the same attention from the stand-
point of specification and materials control as the stainless steel used in the
primary piping. The 1/4-in. type-304 stainless steel had been purchased
to standard ASTM tubing specifications, but in 30- to 40-ft lengths instead
of the usual 20-ft lengths. The tubing was received and installed without
difficulty. It was given a hydrostatic test after installation and put in
service with distilled water. After approximately three months it was
observed that some of the water drained from the leak-detector system
was badly discolored. An analysis revealed the liquid to contain approxi-
mately 1000 ppm of chloride. Since conditions of operation up to this
point had been relatively mild, it was thought that the chloride might be
removed simply by flushing, and approximately six weeks were devoted to
disassembling the reactor and washing out all detectable indications of
that contaminant.

After discussions with the tubing manufacturer it was concluded that
the chloride had originated from a die-drawing compound which had not
been removed from the inside of the tubes prior to annealing. The presence
of the chloride-containing hydrocarbon caused carbide precipitation at the
grain boundaries during annealing and created tiny caves into which the
chloride penetrated. The pickling and cleaning treatment which followed
did not remove this material; in fact, it was learned that the manufacturer’s
pickling tanks did not accommodate the full length of the tubing, making
1t necessary to pickle by dipping approximately half the tubing at a time.
The net result was that large quantitics of chloride remained inside the
tubing to be leached out later when filled with water.

At the time the stress-cracking damage was discovered late in 1956 the
reactor had been made ready for a series of engineering tests, and for this
reason it was decided to make a brief inspection of the damage resulting
from the chloride contamination before proceeding with the planned ex-
perimentation. This preliminary inspection provided the basis for a
decision to prepare for the replacement of the 259 flanges and the 15,000 ft
of leak-detector tubing in the system. It was further decided that engineer-
ing tests which had been interrupted could proceed for the period of
approximately three months which would be required to procure new
flanges and leak-detector tubing.
374 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION Lcaap. 7

Dismantling of the system was begun in April, and a very careful in-
spection of the 259 flanges was made to determine how many should be
replaced. The Super-Zyglo dye-penetrant method of flaw detection was
chosen as the most sensitive test which could be used practically in the
field.

Of the 259 flanges inspected, 167 were found to be acceptable (i.e., as
good as new), 67 were rejected because of cracks, pits, or other possible
flaws, and 25 were judged questionable. Nearly all of the rejected flanges
were in high-temperature portions of the reactor. While the inspection
method was selected as the best available, it was not judged to be infallible;
e.g., differentiation between mechanieal scoring and corrosion pitting was
not always clear-cut, and any cracks covered by smeared metal resulting
from excessive gasket pressure could not be detected. Hence it was de-
cided to replace all the high-temperature flanges with new flanges or, where
this was not possible, to remove 0.02 in. of metal from the flange surfaces.
A total of 132 flanges were replaced; 15 were remachined. In addition,
the 1/4-in. stainless steel tubing to all the high-pressure flanges (approxi-
mately 10,000 ft) was replaced in the leak-detector system. This repair
work was completed in August 1957.

To remove any organic material introduced during repairs and to pre-
treat the fresh metal surfaces incorporated into the system, the reactor
piping was subsequently flushed with hot 39, trisodium phosphate solution,
followed by water rinses and a 5% nitric acid wash. After hydrostatic
testing the reactor was test operated with condensate for 150 hr at 280°C,
then charged with depleted uranyl sulfate solution. Test operation with
depleted uranium included: (1) a series of concentration and dilution
experiments to study the transient and equilibrium behavior of ions in the
system, (2) checking of the inventory-control methods by comparing the
fuel analyses and indicated system controls with the quantity originally
charged, and (3) observation of the corrosion behavior by analysis of fuel
samples during a 159-hr run at temperatures above 250°C. At the conclu-
sion of the run the charge was recovered and found to agree well with the
computed inventory, although chemical analyses of high-pressure-system
samples during operation had indicated a uranium concentration 5 to 10%
lower than the amount added would predict. Nickel analyses of the fuel
solution pointed to a system corrosion rate of slightly less than 1/2 mpy.

Before the reactor was charged with enriched fuel, the piping and shield
were subjected to careful leak tests. To obviate the presence of helium in
the piping in case further testing with helium became necessary, the reactor
was first pressurized to 500 psig with nitrogen, to which was added 40
curies of Kr®? as a tracer gas. The shield was sealed and the reactor allowed
to stand pressurized for five days, after which air samples were drawn from
the shield and heat-exchanger shells for beta-activity scanning to detect
7-4] THE HOMOGENEOUS REACTOR TEST {HRE-2) 375

the presence of any leaking krypton. This test was inconclusive at the time,
however, because of difficulties encountered in purifying the samples for
counting. Large samples of the air being tested were stored in gas eylinders;
then the piping was vented and repressurized with helium. After an addi-
tional waiting period the sealed volumes were again checked using helium
leak detectors sensitive to 1 ppm of helium in air. This test demonstrated
that the piping leakage was less than 0.2 ce/day of helium, with a pressure
ditferential of 15 psi across the leak. These results were subsequently
confirmed by krypton data after analytical difficulties were resolved.

With the integrity of the reactor piping established, the seal pans were
welded in place on the shield roof, and the roof plugs were locked in place.
By pressurizing the shield and flooding the roof with water, leaks in the
seal pan welds were located for subsequent repair. The major shield leak
was found to be through a 16-in. valve in the ventilating duet; when this
was repaired, the total leakage fell from 25 ¢fm to approximately 1,/2 cfm.
The shield was judged sufficiently tight at this point to proceed with
the eritical experiments, after which repairs were continued. By pains-
taking individual checking of all shield penetrations and the use of ther-
mosetting resin to seal the metal lips to which the roof seal pans are welded,
the leakage was finally reduced to 4 to 4.5 liters/min at 15 psig.

7-4.5 Nuclear operation [21]. Fuel charging began on December 24,
and eriticality was achieved on December 27, 1957, with the core and
blanket near room temperature and at a pressure of about 800 psig.
Nuclear instrumentation for the test consisted of three fission chambers,
viz.. two permanent chambers in the instrument tube outside the reactor
pressure vessel and a temporary chamber inside the blanket vessel. An
antimonv-bervllium neutron source was suspended in the thimble in the
center of the core (zee I'ig. 7-18).

The reactor was brought gradually to the eritical condition by the m-
jection of enriched uranium into the fuel solution added to the dump tanks
in batches of 100 to 400 g. Tuel feed pumps and purge pumps were operated
continuously to provide mixing between the dump tanks and the high-
pressure system. Following each addition the solution concentration in the
high-pressure system was allowed to reach steady state, as indicated by a
leveling-off of fission count rates. After 2060 g of U35 had been added
and the temperature of the solution lowered to 29°C, the neutron source
was withdrawn. At this point, the fission count rates continued to rise,
indicating that eriticality had been achieved. Raising the temperature of
the reactor slightly by pumping warm water through the heat exchangers
stopped the nuclear chain reaction. By further varyig the temperature
and concentration of the fuel solution, it was demonstrated that the nu-
clear reaction could be easily and safely controlled in this manner.
376 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

After the initial eritical experiment the neutron source was moved from
the core to the blanket thimble, and the reactor was brought to eriticality
seven times at suceessively higher teraperatures ranging up to 281°C. In
each experiment the reactor temperature was raised above the desired
point by supplying steam to the heat exchangers, and a batch of fuel so-
lution was injected into the dump tanks., After steady count rates showed
complete mixing of the new fuel, the temperature was slowly lowered until
the critical temperature was reached. It was held at this point for about
1/2 hr before procecding to the next experiment. IFigure 7-15 compares
the experimental measurements of the HRIS-2 critical concentration as a
function of temperature with concentrations calculated by various methods.
It can be seen that the two-group caleulations predicted values about 209
below those observed. The harmonics caleulation, which used a convolu-
tion of an age and a Yukawa kernel to represent slowing down in D20,
gave quite satisfactory results. The multigroup ecalculation also gave
results in agreement with the experimental data.

The first operation of the reactor at significant power levels took place
in February 1958. In April 1958 the power level was raised in steps of
I Mw to the design power level of 5 Mw. Operation was exceptionally
smooth, and no mechanical difficulties were encountered in the first 500 hr
after charging the reactor with U??5 Unfortunately, shortly after reaching
full-power operation a crack developed in the tapered portion of the
zirconium core tank, permitting fuel solution to leak into the blanket
After a series of tests to determine the magnitude of the leakage and cal-
culations to determine the behavior of the reactor with fuel in the blanket,
it was decided to operate the reactor as a one-region machine (i.e., identical
fuel solution in core and blanket). Operation of the reactor under these
conditions was resumed in May 1958.

7-4.6 Operational techniques and special procedures. Reacfor starfup.
As the size of homogeneous reactors increases, the use of control-rod neu-
tron absorption to perform a startup becomes progressively less attractive;
e.g., to maintain criticality in HRE-2 while heating from 20 to 280°C re-
quires a reactivity increase of more than 259, Ak because of the large
negative temperature coefficient. It is mueh more convenient to provide a
means of varying the amount of fuel in the core as required to overcome
temperature and power coefficients.

During the initial experimental stage of HRE-2 operation, startup was
begun by evaporating heavy water from the dump-tank solutions, con-
densing, and pumping to the core and blanket circulating loops. The filled
loops were then pressurized, circulation was initiated by starting the
canned-motor pumps, and the circulating stream was preheated to operat-
ing temperature with an auxiliary heat source. The concentrated fuel was
then pumped from the dump tanks to achieve criticality.
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 377

 

      
 
  
 
  
 
  

  

 

 

 

12
l 1 | | i | |
10 [— O Gravimetric ]

% O Apparent (From Multigrqup

o Inventory Data) Calculatien

o

< 8 |
wy
™
N:) Harmonics

o Calculation

c 6 — ]
2

s

£ With Thimbte

e In Core

& 4 |
o

U

o

o

T \ Two-Group Calculation

S 2 = B

i | | l | | |
0 40 80 120 160 200 240 280 320

Temperature, °C

Fi1c. 7-15. Critical concentration of HRE-2 as a function of temperature.

An alternate procedure involves starting the reactor without preheating
hy varving the fuel concentration. When sufficient fuel has been added to
raize the reactor temperature to its normal operating level, power with-
drawal 15 begun.  Temperature adjustments may be made by removing
pure =olvent to temporary storage tanks or adding pure solvent to the
cireulating fuel solution.

A vuriation of this procedure is to fill the reactor slowly with fuel of the
final coneentration. In this case the reactor will become chain-reacting at
a low temperature with the core tank only partially filled with fuel solution.
As the quantity of liquid in the core is slowly increased, the temperature
will rise until the desired temperature is attained with the core completely
full.

The time required for startup is determined by the rate at which heating
can be permitted. The same limitations apply to homogencous reactors as
to other reactors in this respect. Generally, the heating rate of 100°F/hr
is considered reasonable and unlikely to produce excessive stresses. Prob-
ably more important, but more diflicult to determine, 18 the temperature
difference which exists across heavy walls. Keeping these temperature dif-
ferences to less than 100°F prevents excessive stresses. Once the tempera-
ture limitations are established, the startup rate of fuel addition can be
set to mateh them.

Thus, although aqucous homogeneous systems have been demonstrated
to be inherently stable, restrictions are nevertheless placed upon the oper-
378 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cuar. 7

ator to prevent excessive power surges and the resultant excessive pres-
sures or heating rates. These restrictions are generally in the form of elec-
trical interlocks in the control circuit. IFor example, a typical interlock
might prevent the operator from concentrating fuel if instruments indi-
cated that an excessive concentration was being reached, or an interlock
might stop the addition of fuel solution if the temperature-measuring de-
vices ndicated too high a heating rate. Although practically none of these
mistakes Is serious enough to cause a reactor aceident, they are evidenee of
poor operating technique, and if permitted might result in more serious
mistuakes which could cause damage in spite of the inherent stability of an
aqueous homogeneous system.

The reactor 15 considered “'started up” when the desired operating tem-
perature hus been achieved and the reactor is ready for power extraction.

Operation.  LExcept for several preliminary preparations, such  as
warming-up the steam turbine, power extraction is the simplest part of
the reactor startup routine. This involves merely turning steam to the
turbine, bringing the turbine up to speed, and making the necessary elec-
trical switching changes to distribute the clectrical output. Once the gen-
erator is synchronized and feeding into a larger power network, there is
little for the operator to do except to see that the equipment is checked
routinely for proper performance. For HRE-2 this normally takes a crew
which consists of a supervisor, an assistant engineer, and nontechnical
helpers. Checks of all continuously operating equipment are made at 1- to
2-hr intervals to verify that the equipment is performing properly. From
time to time samples of fuel solution are removed from the high-pressure
circulating system to analyze for nickel and other unwanted ions. Samples
must also be removed from the steam system to show that boiler feedwater
treatment is adequate and that oxygen production is not excessive in the
steam generators. Radiation levels must be observed to determine whether
there are leaks of fuel solution or weak places in the shield structure.

In addition to these routine service functions, the operating crew must
also start up and maintain the chemical processing plant associated with
the reactor. One engineer and one technician, in addition to those already
mentioned for the reactor proper, are required to operate the HRIE-2
chemical plant.

Shutdown. The shutdown of the reactor is normally accomplished in
two steps.

The first step in the procedure is cessation of power removal, which in-
volves nothing more than closing the steam throttling valve. This might
be accomplished by running the turbine governor down to zero load.
Although this action leaves the reactor critical at 280°C and does not com-
pletely stop heat generation, the power is limited to the normal heat losses
from the system (a few hundred kilowatts for HRE-2). Minor repairs or
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 379

adjustments to equipment which do not require further cooling can now
be made.

The second step is to make the reactor suberitical by diluting the fuel.
This type of shutdown normally requires 3 to 5 hr, since it is desirable to
have the reactor subcritical at the storage temperature of approximately
25°C when the reactor is emptied. The rate at which the temperature can
be reduced is again determined by the permissible cooling rate of the
system components.

Scram. A more rapid shutdown, equivalent to an emergency scram in a
solid-fuel system, is the “‘dump.” In this situation the reactor is kept circu-
lating for 2 min to permit recombination of the radiolytic gas in solution
during which time the steam valves are closed to reduce power output.
Then the pressure in excess of the vapor pressure of the core and blanket
is vented to make pressure balancing between core and blanket easier,
and the dump valves are opened, permitting the rapid emptying of the
core and blanket vessels so that the reactor is shut down within minutes.
This type of shutdown is only resorted to in case of emergencies such as
excessive pressures or evidence of a leak of radioactive solution from the
reactor.

Decontamination of equipment. Conventional methods of decontaminat-
ing fuel processing plants have proved to be inadequate for a stainless steel
homogeneous reactor system which has been exposed to uranyl sulfate-
sulfuric acid-fission product solutions at 250 to 300°C. This was demon-
strated with HRIE~1, which was decontaminated (without descaling) prior
to disassembly (Article 7-3.13). Since the HRE decontamination was in-
complete, laboratory studies were carried out to explore the nature of the
contamination and develop methods of decontaminating the stainless
steel. Tt was found that chromous sulfate, a strong reducing agent, would
modify the oxide film and permit dissolution in dilute acids. A 0.4m
CrS04+—0.5 m Ho80,4 solution has given excellent removal of the film by
modifying and dissolving the oxide corrosion film. Decontamination fac-
tors of 5 X 10° were achieved on specimens from in-pile corrosion loops,
where the activity was reduced to the induced activity of the structural
material, by contacting for 4 hr with the chromous sulfate solution at 85°C.

The solution was also tested satisfactorily on four 22-liter uranyl sulfate
corrosion loops which had run for 22,000 hr at 200 to 300°C. The loops
had a very heavy oxide coating such that in thermal cycles large flakes
broke off the wall and plugged small lines. After a 4-hr contact with the
chromous sulfate-sulfuric acid solution at 85°C, the walls of the loop were
completely free of all clinging oxide.

The total time involved in the preparation of the chromous sulfate
solution, in descaling the reactor system, and in disposing of the extremely
radioactive scale-waste would probably be at least 48 to 72 hr. If one is
380 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

considering decontamination as an aid to maintenance, time required for
decontamination should be weighed against the reduction in repair time
which would result from the lower levels of activity in the working area.

7-4.7 The HRE-2 Mockup [22].* During the design and construction
of HRE-2 some of its major pieces of equipment were assembled and
operated at design conditions on unenriched uranyl sulfate solution. The
purposes of this engineering mockup were (1) to study the behavior and
removal of gases in the high-pressure system, (2) to study fuel-solution
stability in a circulating system similar to the HRE-2, (3) to establish a
reference corrosion rate for the reactor circulating system, (4) to study the
behavior and removal of corrosion- and fission-product solids in the system,
and (5) to establish the reliability of components, pointing out weaknesses
and possible improvements.

The prototype reactor components tested in the mockup include: (1) a
Westinghouse canned-rotor centrifugal pump, (2) an electrically heated
steam pressurizer, (3) a centrifugal gas separator, (4) a 1/8-scale heat
exchanger similar to the HRE-2 steam generator, (5) a letdown heat ex-
changer to cool the fluid and gas bled from the high-pressure system,
(6) a bellows-sealed letdown valve to throttle the fluid and gas from the
high-pressure system, (7) a liquid-level controller which adjusts the
letdown-valve position, (8) the dump-tank and condensate system for
excess fuel solution and storage of condensate for use as purge to the
pressurizer and circulating pump, (9) diaphragm feed pumps to return
fuel to the high-pressure system from the dump tanks, (10) an oxygen-feed
system to maintain oxidizing conditions in the eirculating fuel, and (11) an
air-injection system for tests of the effectiveness of the gas-separator unit
and letdown system.

In May 1956 a corrosion- and fission-product solids-removal system was
mstalled m the mockup, consisting of a 5-gpm canned-rotor ORNL pump,
an assembly of a hydroclone and underflow pot, and two 1/2-in. I'ulton
Sylphon bellows-sealed air-operated valves separating the system from
the main circulating stream. A commercial pulsafeeder was used to inject
rare earths containing radioactive tracers and corrosion products. In ad-
dition, a through-flow bomb was filled with the required solids and con-
nected into the system. Gamma-ray counting equipment was installed to
detect any buildup of radioactive tracers on the heat-exchanger surfaces,
in the horizontal connecting pipe to the pressurizer, in the gas separator,
and in the underflow collection pot. A multichannel gamma counter was
used for determining activity levels at the various counting stations. The
equipment was removed in November 1956 after demonstrating satis-
factory removal of solids from the system.

*Article 7-4.7 is based on a paper by L. Spiewak and H. L. Falkenberry [22].
7-4] THE HOMOGLNEOUS REACTOR TEST (HR-2) 381

Operation of the mockup for a total of more than 13,000 hr in the three-
year period from Iebruary 1955 to February 1958 provided valuable in-
formation pertinent to the construction and operation of HRE-2. Items
of major importance include: (1) a satistactory demonstration of the opera-
tion of the equipment for removing gases from the high-pressure system,
(2) the out-of-pile chemical stability of the uranyl sulfate solution over a
long period of time at operating temperatures and pressures, (3) determina-
tion of the oxygen injection and excess sulfuric acid requirements to pre-
vent uranium precipitation in the high-pressure loop, (4) detection of an
unsatisfactory pressurizer design in which excessive corrosion and precipi-
tation of uranium occurred, and tests of a revised model which proved
suitable, (5) & demonstration of the suceessful removal of injected fiscion-
product solids and insoluble corrosion products by means of a hydroclone,
and (6) long-term operability of the circulating pump and other pieces of
equipment,

7-4.8 The HRE-2 instrument and control system [23].* The control
system for & homogeneous reactor such as the HRIS-2 differs drastieally
from that for solid-fuel reactors hecause control rods and fast electronie
circuitry are not necessary for systems with such large temperature co-
efficients (nearly 0.9, Al/°C at 280°C for the HRIS-2).

I'unctions similar to those performed by control rods in heterogeneous
reactors, but without such exacting speed-of-response requirements, are
performed by valves which control the concentration of the fuel, which
ary the steam-removal rate from the heat exchangers, or which allow the
fuel to be discharged to noneritical low-pressure storage tanks, Ifor these
rensons the nuclear control eircuits for a homogeneous reactor are designed
to limit very rapid changes in fuel concentration and steam-removal rates.
Other circuits control the pressures und temperature limits of the cireulating
liquid fuel, principally to prevent equipment damage.

In addition to these general considerations, the instrument and control
system for the HRE-2 includes the following special features:

(2) All instrument lines through the shield wall are blocked by valves
on a signal of high shield pressure to prevent the escape of radioactivity.
Electrical leads are sealed by glass-to-metal seals.

(b) Thermocouple, electrical, and air lines may be disconnected by
remotely operated tools for equipment repair or removal.

(¢) All eritical core- and blanket-system transmitters, except electric
level transmitters and thermocouples, are located in two shielded nstru-
ment cubicles (5 ft diameter X 15 ft high) located adjacent to, but out-
side, the main reactor tank. This arrangement wus selected to avoid
opening the main tank to replace instruments, to provide a location for

*Article 7-4.8 is based on a paper by D. 8. Toomb [23].
382 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cmAP. 7

 

 

   

 

Pressure
Switches Pressure
- Switches
Eiectric
Controller }-{ Recorder Pneumatic
Converter

 

 

   
  

 

 

 

 

 

Control
Circuit

 

 

 

 

Dump
Tank

Pulsafeeder
Pump

 

Fia. 7-16. Key control loops utilizing both pneumatic and electric transmission.

instrument components which could not be easily protected from the
water-flooding of tne main tank during remote-maintenance operations,
and to minimize the radiation exposure of instruments.

(d) The cell air monitors, which provide an alarm in case of a leak of
radioactive vapor from the reactor system, are installed in one of these
mnstrument cubicles. Cell air is circulated through a 2-in. pipe from the
reactor tank, past the enclosed monitors, and then back to the cell. The
blower is sized so that only 5 sec is required for cell air to reach the radia-
tion monitors.

I'igure 7-16 indicates several key control loops: (a) The pressure of the
core system 1s controlled from sensed pressure by the proportioning of
power to the pressurizer electric heaters; the blanket pressure is similarly
controlled by a core-to-blanket differential-pressure signal.  (b) The
liquid levels in the pressurizers are controlled from sensed levels by pneu-
matic control of the letdown valves. Pneumatic control actions are derived
from transducers which receive signals from electric transmitters. Electric
interlock control of the pneumatic signals to final control elements is
achieved by the use of solenoid-actuated pilot valves.

Other control loops not illustrated are: (¢) The reactor power is con-
trolled by a manual or turbine-governor signal to a valve throttling steam
from the core heat exchanger. Because of the large negative temperature
coefficient and low heat capacity, this reactor system cannot produce more
heat than is being removed, except for very short times. Heat is removed
by opening the steam throttling valve, which causes a decrease in the
temperature of the fuel leaving the heat exchanger. Because of the nega-
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 383

Overflow Pipe

   
  
 

To Wcsie/

 

Reactor Thermal Shield

 

  
  
 

Boral Sleeves

 

12#0in, ———=

L |

Fia. 7-17. HRE-2 nuclear instrument thimble.

 

 

 

tive temperature coefficient of the system, the cooler fuel entering the
core causes an increase in reactivity and the fuel is reheated until it over-
comes the excess reactivity. (d) The average temperature of the core
system is controlled by varying the concentration of the fuel solution.
The inlet and outlet temperatures will vary with the power extraction,
while the average temperature is a function only of fuel concentration.
(e) The blanket temperature is controlled by a signal, derived from the
difference in average core and blanket temperature, which operates the
steam withdrawal valve of the blanket heat exchanger.

Nuclear instrumentation. Neutron level transmitters are two Westing-
house fission chambers and two gamma-compensated ilonization chambers
designed by the Oak Ridge National Laboratory. Neither type of instru-
ment requires a gas purge, and both are amenable to operation in the water-
filled thimbles (I'ig. 7-17), which allow the chambers to be positioned or
replaced during reactor operation. Varying the distance of the chambers
from the reactor core affords a means of sensitivity adjustment, which is
needed to accommodate different operating powers.

The gamma-compensated ionization chambers are required to be in a
384 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHaPp. 7

neutron flux of approximately 10'¢ at 5-Mw power to utilize their measur-
ing range of 10°. The lower limit of their range of operation is comparable
to the maximum desirable flux for proper operation of the fission chambers,
Therefore 1t was possible to install the two types of sensing elements in
the same area in the reactor cell. At high fluxes the fission chambers which
are required to follow the neutron flux for five decades during start-up are
withdrawn into a protective boral shield to limit fission-product buildup
in the chamber lining. Proper operation of the compensated ion chambers
is ensured by the lead-shot-and-water fill around the thimbles, which re-
duces the gamma-ray background after power operation from 250,000 r/hr
to 250 r/hr.

The fission-chamber signals are fed to conventional preamplifiers,
A-1 linear amplifiers, logarithmic count-rate meters and a dual-pen re-
corder. For initial reactor startup before gamma-ncutron reactions pro-
vided a sizable neutron source, the fission-chamber output was used to drive
a low-range pulse counter.

Control panel. Figure 7-18 shows the main control board and console.
Here are located only those instruments necessary for the safe operation
of the reactor. These are arranged in a “visual aid” form to reduce opera-
tional errors and to facilitate the training of operators. The graphic section
is essentially a simplified schematic representation of the chemical process
flowsheet with instruments, control switches, and valve-position indicators
located In positions corresponding to their location or function in the actual
system.

Each annunciator is placed in the control board directly over the instru-
ment or portion of the system on the graphic board with which its signal
is associated.

Key measurements are displayed on “full-scale” recorders in the center
section of the panel and include the fuel temperature, a multipoint tem-
perature recorder, the multiarea radiation monitoring recorder, the re-
actor power, the logarithmic neutron-level and count-rate meter signals.

The patch panel on the extreme left 15 a “jumper board,” which is a
schematic representation of the electrical control circuits. Provision is
made for jumping certain individual contacts in the control circuit with a
plug. Lights indicate the position, open or closed, of the contact in the
system. The board is valuable for making temporary control-circuit altera-
tions necessary for experiments and makes 1t unnecessary to use jumper
wires which are normally placed behind the panel and if overlooked and
not removed might introduce a hazardous condition. The board is also an
aid in familiarizing operators with the electrical control circuitry and, since
the lights indicate contact position, as an operations aid during startup.

Switches and controls on the console are restricted to those necessary
for nuclear startup, steady-state power operation, and emergency.
THE HOMOGENEOUS REACTOR TEST (HRE-2) 385

7-4]

"[Pued JUIWNIISUI PUB WOOI [013U0d G- H 'S1-L D1

 
3806 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHaP. 7

Data-collection instruments and the transducers that drive the miniature
pneumatic slave recorders on the graphic panel are located in an auxiliary
instrument gallery beneath the main control room, as are a 548-point
thermocouple patch panel, a relay panel, the nuclear-instrument amplifiers,
and the nuclear-instrument power supplies.

Other panels located near their respective equipment elsewhere in the
building include the steam control station, the turbine control panel, two
sampler control panels, and a control station for the refrigeration system
supplying the cold traps. Standard 2-ft-wide modular cabinets and panels
are used throughout to facilitate design changes.

Protective interlocking. Extensive protective interlocking of the controls
circuits is provided to prevent unsafe operating conditions.

Examples of interlocked systems include the following: (a) The
pumping-up of fuel to the reactor core instead of condensate is prevented
by several interlocks which keep the fuel-addition valve closed until the
core is full of condensate and has been heated to 200°C. This prevents
power surges and consequent pressure increases. (b) For the same reason
the fuel circulating pump is started in reverse to provide a low flow rate as
protection against pumping cold fuel too rapidly from the heat exchanger
into the core. (¢) Toavoid dangerous thermal stresses and abrupt reactivity
surges, the control circuits do not permit the pumping of cold feedwater
into the heat exchangers until the level is above 509%. (d) To give smooth
startup, the fuel injection pump can run only at half-speed until a tem-
perature of 250°C is reached. {(e) The fuel feed valve will be closed and the
concentration in the fuel system will be lowered by injecting condensate
if the core outlet temperature becomes excessive, if the circulating pump
stops, or if the power level exceeds normal. (f) The contents of the high-
pressure systems will be automatically emptied to the low-pressure storage
tanks through the “dump” valves on a signal of extremely high pressure,
or a radiation leak into the steam system. Differential-pressure control
between the core and blanket systems during dumps is by throttling control
of valves from a differential-pressure signal.

Inventory systems. To obtain an accurate inventory of the fuel and
moderator solutions, the storage and condensate tanks are weighed with
pneumatic weigh cells. A pneumatic system was selected primarily because
taring can be done remotely by balancing air pressures, and components
are less susceptible to radiation damage. Piping to the tanks is kept
flexible to compensate for the varying loads which result from thermal
expansion.

The volumes of the fuel and blanket high-pressure systems have been
accurately measured, so that when the pressurizer condensate reservoirs
are filled to capacity, the weight of liquid in the high-pressure systems can
be computed from the core, blanket, and pressurizer temperatures. These
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 387

weights, combined with the respective condensate weights, dump-tank
weights, and experimentally observed holdup in the condensate piping,
yield the total liquid inventory of the reactor.

7-4.9 Remote maintenance [24]. Maintenance of the equipment in the
circulating systems of an aqueous homogeneous reactor is difficult because
of the intense radioactivity emanating from the surfaces following pro-
longed operation at high power levels. This problem, more acute than in
the case of heterogeneous reactors because fission-product activities and
neutrons are not confined to the reactor core, must be successfully resolved
in a practical manner if the homogeneous reactor is to be an economical
power producer. Practical systems for maintenance and repair must pro-
vide adequate personnel shielding to reduce radiations from activated
equipment to tolerable biological values while simultaneously providing
access to the equipment.

Equipment activation. Of the two principal methods by means of which
the equipment is rendered radioactive, activation by fission-product con-
tamination is the more important in the HRE-2. About 30 w/o of the
fission produects is normally removed from the fuel solution as gases (xenon,
krypton, iodine) by stripping with Oz, D2, and steam. Of the remaining
fission products, 56 w/o are estimated to have sufficiently low solubilities
to precipitate from the solution as solids. This is particularly true of the
rarc earths, which constitute 28 w/o of the fission products. Tellurium,
technetium, and molybdenum, which contribute 12 w/o of the fission
products, are assumed to be soluble. Therefore, so far as activation of
equipment is concerned, the rare gases { ~309) and the soluble products
(~12%), making up 42 w/o of the total, may be considered removed or
easily removable from the equipment. Some portion of the remaining
58 w/o will deposit on hot metallic surfaces or be retained in cracks and
crevices. It is possible by statistical analysis to estimate within a factor of
two the specific gamma and beta activity of the fuel solution. Methods for
caleulating this activity have been published [25,26]. The 1nitial fuel ac-
tivity, after prolonged HRE-2 operation, is of the order of 25 curies/ml.

It should be noted that it is proposed to remove the insoluble fission
products continuously with small centrifugal-type separators (hydroclones)
in the HRIE-2 fuel processing system. The effectiveness of this procedure
in competition with absorption on the container walls, however, is us yet
unknown,

Induced activation of structural materials by neutron absorption ig also
an important contributor of radioactivity. During operation of the reactor,
a thermal flux exterior to the equipment items will be present because of
neutron leakage from the reactor vessel. Also, delayed neutrons will be
emitted in the interior of the piping and equipment from fission products
388 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

in the fuel solution. These neutron fluxes may be estimated, and knowing
the surface areas and material constituents of the equipment, the resulting
induced activation may be calculated. Methods of calculating induced
activation are reported in the literature [27,28].

The effect of induced activation may be controlled to some extent by
attenuating neutron leakage from the reactor through an adequate thermal-
neutron shield surrounding the reactor vessel. Also, the constituents of
the structural materials in the cell may be specified in such & manner that
elements of potentially high neutron activation, such as the cobalt normally
present in stainless steels, are minimized. Other materials, such as boron,
which has a high neutron-capture cross section and is an alpha emitter
rather than a gamma radiator, might be used on equipment surfaces to
minimize induced activity.

Radiation dosages. Design criteria followed in the design of shielding and
tooling for remote-maintenance operations are based on tolerance radia-
tion levels that are acceptable for normal operations. In continuous work-
ing areas, radiation levels as low as 1 mr/hr are specified. For certain in-
frequent operations a level of 74 mr/hr is permitted. However, during
periods of intense maintenance it may sometimes be permissible to allow
personnel to work in an area of higher than normal radiation, but the total
weekly radiation dosage must be held to 300 mr or less. In some instances
it is necessary to work in relays so that no one person receives a higher than
permissible weekly dosage.

Proposed maintenance concepts [24]. The concepts which have been pro-
posed for the maintenance of equipment in the HRIE-2 have been based
on the philosophy of either dry maintenance or underwater maintenance or
combinations of the two. In general, it is assumed that the item requiring
repair will be replaced with a new item and the faulty component removed
to a shielded hot-cell area for maintenance,

Tests of underwater-maintenance procedures proposed for HRE-2 using
special tools and equipment have demonstrated the feasibility of removing
and replacing equipment. In these tests it was demonstrated that con-
tamination of the heavy-water fuel solution by shielding water could be
prevented by freezing a plug of heavy ice in the pipe before disconnecting
equipment. Ireeze jackets are placed around piping at flange disconnects
for this purpose. Also in the HRIE-2, the insulation around piping was de-
signed to avoid the retention of water after draining the flooded equip-
ment.

Piping joints. Removal of equipment in the HRE-2 depends on flanged
joints adjacent to the equipment which may be disconnected and remade
with suitably designed tooling. The joints in the HRE-2 use American
Standard stainless-steel ring-joint, weld-neck flanges modified to permit
leak detection (Fig. 7-13). These flanges are located so that all bolting is
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 389

 

Fig. 7-19. Universal jig for locating flange positions on HRE-2 replacement
high-pressure valves.

accessible to tools operated from above. The number of loose pieces such
as bolts, nuts, ferrules, and gaskets is minimized by welding nuts to the
flanges such that they become integral with the flange.

Remotely operable tools.* Tools used in the IHIRE-2 maitenance pro-
cedures are generally simple, rugged units mounted on long handles.
They include manual, air-operated, and water-operated socket wrenches
for turning nuts, hooks for lifting, clamps, knives, magnets, etc. Some
typical maintenance devices are illustrated in Iligs. 7-19 through 7-22
and described as follows.

The universal high-pressure valve jig shown in Fig. 7-19 is used for
manufacturing replacement high-pressure flanges for the HRIE-2. By
placing brackets in appropriate holes, referenced with respect to the 21
high-pressure HRE-2 valves, the position of the flanged pipe ends for any
one of the valves may be accurately determined. With this device, the re-
quired number of spare valves for the reactor is minimized, and at the
same time any valve which fails can be remotely replaced.

*Other remotely operable tools are described in Sec. 19.5.6 of the Reactor
Handbook, Vol. II, Section D, Chapter 19, ORNL-CF-57-12-49.
390 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

 

F1a. 7-20. Spinner wrench for rapid removal of HRE-2 flange bolts.

The spinner wrench in Fig. 7-20 is used for removing bolts quickly after
they are loosened with a high-torque wrench, for spinning the bolts snug
before final tightening and, by use of a right-angle bevel-gear attachment
(not shown), for handling holts in a horizontal position. It has a chain-
drive offset to fit bolts not accessible by direct drive, such as bolts hidden
under pipe fittings and bends.

The flange-spreader tool (Fig. 7-21) is required in remote operations to
spread low-pressure valve flanges to prevent damage to the ring gasket
during removal or ingertion of the valves. The tool is shown in place with
the flanges spread and the valve assembly moved back on its cell fixture.

The hydraulic torque wrench in Fig. 7-22 is used to tighten or loosen
bolts on flanges larger than 2-in. pipe size. This wrench is actuated by a
hydraulic (water-operated) cylinder and is capable of applying a 2000 ft-lb
force moment.

Even in the relatively short operating history of HRIS-2 the concept of
underwater repairs with tools such as those described above has been
proved to be completely practical. Soon after the reactor attained signifi-
cant power it was found necessary to replace all the clectrical power wiring
—nearly 4000 ft of wire. This was accomplished in & period of three weeks,
with the cell completely flooded. During this period the fuel circulating
pump was removed to recover a set of corrosion samples. The removal
and replacement of the pump required only 100 hr working time, which
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-Z) 391

 

Fi1e. 7-21. Flange-spreader tool for preventing damage to gaskets during re-
moval of low-pressure flanges. Shown in place during maintenance operation on
HRE-2.

would have been the total shutdown time if the pump replacement alone
had bteen performed. Also, the core inspection flange was removed and the
core mner surface inspected by means of a periscope. In these three major
underwater-maintenance operations no delay or difficulty was experienced.

7-4.10 Containment methods [29]. In designing for the containment of
HRE-2, extremely rigid leakage specifications were set, both for the pri-
mary piping system and the shielded tank containing the reactor. In the
case of the piping, the leakage specification was based on the minimum
leak which could be found with the available mass-spectrometer leak-
392 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

 

N

Fic. 7-22. Hydraulic torque wrench for loosening or tightening flange bolts.
Shown in operation on HRE-2,

detection equipment, namely, 0.1 cc helium (STP) per day. Allowing for
several such minimum leaks, the equivalent loss of liquid at 300°C and at
2000 psig could be as great as 5 cc per day, or a total leakage of approxi-
mately 250 curies/day. Because of its volume and surface area, and be-
cause of the difficulty in measuring small leakage from very large vessels,
the leakage rate for the reactor tank was set at 10 liters/min or less, at a
test pressure of 15 psi. As indicated in Article 7—4.4, actual leakages were
below this value.

The problem of leakage from the reactor vessels and piping can be cata-
logued according to the various mechanisms through which leakage might
result as follows: (1) excessive stresses, (2) defective materials or work-
manship, (3) corrosion, (4) nuclear accidents, (5) hydrogen-oxygen ex-
plosions, and (6) brittle fracture. Each possibility is examined in detail in
the discussion which follows.

Ezxcessive stresses. There are many possibilities for the development of
excesslve stresses in a system as complex as the HRE-2. In order to reduce
the likelthood of failure as a result of excessive stress, a maximum allowable
working stress of 12,000 psi was specified for the type-347 stainless steel
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 393

with which the system was fabricated. This permits an additional factor
of safety over the 15,000 psi allowed by the ASME boiler code. As required
by the ASA code for pressure piping, the reactor piping arrangement was
examined for maximum stresses due to pressure, as well as for hoop and
bending stresses resulting from thermal expansion. Equipment was also
studied for determining the magnitude of thermal stresses caused by radia-
tion heating and temperature cycling. Therefore, in order to keep the com-
bined thermal and pressure stresses below the maximum allowable working
stress, the pressure-vessel wall is approximately 2 in. thicker than would
be required otherwise. Heating and cooling rates on the entire system have
been limited to 100°F/hr and 55°F/hr, respectively, and the differential
temperature across heavy metal walls is kept below 100°F. (The cooling
rate is held below the heating rate to minimize rate of flange contraction.)

Cyclic temperature stresses at questionable points in the reactor pres-
sure vessel and steam generators were explored experimentally. Mockups
were fabricated for the testing of the pressure-vessel nozzle joints and the
stainless-steel-to-Zircaloy bolted joint inside the pressure vessel. In each
test the temperature was cycled from approximately 250°F to approxi-
mately 600°F in 1/2 hr and cooled back to 250°F in 1/2 hr. After 100
cycles the joints were found to be sound. The main steam generators were
also cycled in similar tests. Several tube joints cracked open during the
first 50 cycles. They were repaired and the heat exchangers were subjected
to an additional 10 cycles before final acceptance.

Defective materials or workmanship. Defective materials and poor work-
manship constitute another area which required special attention to pre-
vent failures. All materials for the primary systems of HRE-2 were
procured to specifications considerably more rigid than those existing in
commercial practice. Optional requirements such as chemical analyses,
boiling nitric acid tests, and macroetch tests were exercised in all materials
specifications. An additional cost, averaging about 10%, was experienced
in the purchase of materials under the more rigid specifications. In some in-
stances, for example with the heat-exchanger tubes, special ultrasonic and
magnetic eddy-current flaw detectors were employed to indicate defective
parts. Three tubes which might have later failed in operation were thus
eliminated. Dye-penetrant tests were applied to tubing bends and to all
welds throughout the reactor to detect eracks and pinholes. None was dis-
covered in tubing bends, but many were found in welds, especially in the
tube-to-tube-sheet welds.

Special attention was given to the welding of stainless steel butt joints,
of which there are approximately 2000 in the entire reactor. The inert-gas,
nonconsumahle-electrode method was used almost entirely. Welds were
inspected to considerably higher standards than required by the ASME
code. In addition to being subjected to dye-penetrant inspection, every
394 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

weld was x-rayed. Although the inspection standards were very rigid, only
39 of the welds were rejected, necessitating rewelding.

Corrosion. Although corrosion is an ever-present possibility for leakage,
the HRE-2 was designed to reduce attack rates so far as possible by keep-
ing velocities below 20 fps, temperatures in the range of 250 to 300°C, and
lining some surface areas with titanium for additional resistance.

Nuclear accidents. Nuclear accidents are likely to be rare in homogeneous
reactors because of the large negative temperature coeflicient (0.1 to 0.2%
Ak/°C) [30]. As an example, the worst aceident considered in the HRE-2
[31] was one in which all the uranium suddenly collects in the reactor core
and results in a reactivity increase of 2.5% Alk/sec. For this rate, starting
at a power of only 0.4 watt, the maximum pressure in the pressure vessel
would be approximately 3900 Ib, and the pressure stress in the carbon steel
shell would be less than 30,000 psi.

Hydrogen-oxygen explosions.  Since radiolytic gases (deuterium and
oxygen) are produced continuously in the reactor, they are the source of
an ever-present hazard. Explosions may be expected whenever the
deuterium-oxygen-steam mixture is more than 15% gas. For detonations
the required gas fraction is greater. The maximum increase in pressure
from an adiabatic explosion of hydrogen and oxygen is only a factor of
3 to 8, whereas for a detonation the factor might be 23 for an undiluted
mixture. (It is important to note that detonations can occur only In gas
channels that are relatively long and straight.)

In low-pressure areas of HRIZ-2 the gas was diluted with steam to keep
it noncombustible. Furthermore, a pressure of 500 psi is the basis for
design of low-pressure (atmospheric) equipment. Even for a detonation
the expected peak pressure would be less than the design pressure.

Although an explosion can be tolerated in the high-pressure system with
little danger of vessel rupture, a detonation probably could not be. A
detonation could occur in the gas separator where the gas channel (the
vortex by which the radiolytic gas is collected) is long and straight. It is
caleulated [32] that a detonation wave traveling longitudinally along the
vortex would produce impact pressures of the order of 30,000 psi but that
the damage would be limited to the directional vanes inside the separator.
Attenuation of the forces by the solution would limit the pressure rise to
that resulting from combustion of the gas—10,000 psia, which produces a
tolerable wall fiber stress of 35,000 psi. Thus no serious damage is foreseen
from explosions or detonations.

Brittle fracture. 1t is generally known that ferritie steels are subject to the
phenomenon of brittle fracture. Although the likelihood of brittle fracture
in the HRI-2 pressure vessel was known to be small, an investigation was
made [33] in order to determine the consequences of such an accident as
a result of the pressure rise and missile damage.
7-4] THE HOMOGENEOUS REACTOR TEST (HRE-2) 395

A sudden release of the liquid contents of the fuel and blanket systems—
3910 1b of solution at 300°C and 2000 psig—would result in a pressure rise
of approximately 30 psi inside the reactor shield. Although it is somewhat
difficult to design a large (25,000 cu ft) rectangular container such as the
HRE-2 container to withstand a 30-psig pressure, it was even more diffi-
cult to design it to withstand heavy missiles. Studies indicated that missile
velocities of 50 to 150 fps could be expected. With the HRE-2 vessel,
which weighs 16,000 1b, 1600-1b fragments might be expected. Such a mass
at 100 fps would lift an unrestrained 5-ft-thick concrete shielding plug
nearly 1 ft. For this reason a blast shield was placed around the vessel.

The blast shield was designed to withstand the 1250-psig pressure which
would result from the 300°C liquid, as well as to absorb the energy of the
expanding steam. A 13-in.-thick wall made of type—304 stainless steel was
found to be eapable of absorbing 2.85 X 10° ft-1b of energy with 29 elonga-
tion. The fragments from the pressure vessel would accumulate this much
energy in traveling across an annulus of 4.8 in. To provide an additional
factor of safety, the blast shield was constructed to surround the pressure
vessel with an annulus not greater than 2 in.

A similar shield made of carbon steel was installed on both fuel and
blanket steam generators as a final precaution against an explosion which
might damage other equipment and the vapor barrier sufficiently to cause
a release of activity. The probability of failure of the steam generator
shells would not be affected by the presence of neutron radiation because
the generators are located in a low flux region.

The vapor-tight container. If leakage of radicactive solution from the
reactor proper occurs through one of the means mentioned above, an all-
welded, vapor-tight tank provides a second barrier to the escape of radia-
tion to the atmosphere. Construction details of this tank, which forms the
liner for the biological shield, are illustrated in Fig. 7-8. After completion
of construction the tank was given a hydrostatic test at an average pressure
of 32 psig, and the welded joints were found to be free of leaks. It was then
reopened for further installation of reactor equipment. Prior to operation
of the reactor at power, the containment vessel was sealed and tested
(Article 7-4.4).

7—4.11 Summary of HRE-2 design and construction experience. Fol-
lowing the completion of HRE-2, personnel associated with the project
recommended design improvements which might be applied to good ad-
vantage in future homogeneous reactors. These recommendations are
summarized in the following paragraphs:

(1) There are many components in the HRIE-2 system which must
operate in conjunction with one another for continuous operation of the
reactor. Since a failure of any one of these could cause a shutdown of the
396 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

reactor, the probability of shutdown is higher than it would be with com-
ponents operating independently. In future reactors an attempt should be
made to decrecase the dependency upon the simultaneous operation of
essential components. An example of this is a system which can be operated
without continuous letdown from the high-pressure to the low-pressure
system, so that failure of the low-pressure system components will not
necessarily require shutdown of the entire reactor system.

(2) The reactor cell for the HRI-2 contains many structural compo-
nents made of carbon steel. I'looding the cell with water and removing
radioactive contamination by use of acids will eventually do serious damage
to all these carbon steel surfaces. Wherever possible in future reactors,
corrosion-resistant materials should be used for structural components, or
those components should be protected by coatings which are resistant to
radiation, as well as to acids and water.

(3) In laying out equipment and piping, provision should be made for
walkwayvs, stairs, and ladders to provide convenient access during con-
struction without having to elimb on the process and service piping. These
walkways, ladders, and stairs may or may not be removable after com-
pletion of construetion.

(4) Many items installed inside the reactor cell of the HRE-2 might
have been located outside the radioactive area. These include the cold
traps, which could be located in one of the shielded waste-system compart-
ments, the space coolers, and a number of other nonradioactive components.

(5) In the IHIRT-2 the sampler 1s located adjacent to the cell shield, and
it is necessary to remove the sample from the sampler and carry it to the
analytical laboratory in a shielded carrier. The risk of contamination can
be greatly reduced i the hot laboratory is built around the sampler so that
samples may be analyzed without removal from the shield; however, this
would add to the cost of the facility.

(6) Access through the containment seal, which consists of welded pans,
can be made easier through the use of mechanical seals with organic gaskets.
These would not be subject to severe radiation damage since the coutain-
ment seal 1s itself shielded from primary radiation.

(7) Clontrol of all operations affecting the reactor should be located
in one central control room. This includes the controls for the chemical
processing system, as well as those for the reactor system and all main con-
trols for the steam turbine and steam auxiharies. This does not mean, how-
ever, that all instrumentation must be in the control room, but only those
things which are necessary to keep the reactor operator informed and to
provide him with primary control of the operation.

(8) In the HRE-2 refrigeration system a petroleum distillate, “AMSCO
125-82," was circulated inside the reactor cell as a secondary coolant be-
cause no primary low-pressure refrigerant could be found which would not
7-5] THE LOS ALAMOS POWER REACTOR EXPERIMENTS 397

decompose under irradiation to release chlorides or fluorides detrimental to
the stainless steel piping. Carbon dioxide was not selected because the
piping would have to be designed to withstand about 1500 psi. A solution
to this problem would be to use copper for all tubing and equipment in con-
tact with the refrigerant, in which case I'reon could be used inside the cell.

7-4.12 HRE-2 construction costs. These costs are summarized in Ta-
ble 7-7, including that portion of the costs due to the requirement for pre-
venting the possible escape of radioactive material from the reactor to the
building or surroundings.

7-5. Tur Los Aramos Powrr REACTOR INXPERIMENTS
(Larre I anp 2) [34, 35]*

7-5.1 Introduction. Homogeneous reactors fueled with uranium oxide
dissolved in concentrated aqueous solutions of phosphoric acid have a
number of advantages compared with dilute aqueous uranyl solutions for
certain power reactor applications. These applications are based on the
fact that concentrated phosphoric acid solutions have high thermal sta-
bility and low vapor pressures. This makes it possible to operate at rela-
tively high temperatures without creating the excessive pressures en-
countered with dilute agueous solutions. These temperatures are high
enough to take advantage of the back reaction for recombination of gases
produced by radiolytic decomposition and to eliminate the need for an
external gas system or an internal catalyst. A phosphorie acid solution
reactor, therefore, can be a sealed single vessel with no external compo-
nents except a circulator for the steam system. In addition, the high
hydrogen density and the relatively low neutron-absorption cross sections
of the phosphoric acid system permit the construction of small, compact
reactors with a low inventory of fuel, which may be ideal for remote
package power plants. Other advantages include the possibility of continu-
ous removal of fission-product, poisons and a strong negative temperature
coefficient of reactivity.

Although the advantages of uranium oxide in concentrated phosphoric
acid solutions present a strong case for such a fuel system, the disadvan-
tages, though few in number, are ponderous. In fact, one disadvantage,
the highly corrosive nature of the phosphoric acid to most metals, may well
outweigh all of the advantages. The only metals known to be suitably re-
sistant to phosphorie acid in the coneentration and temperature ranges of
interest are gold and platinum. Although with proper design the cost of
using such materials in a reactor can be kept within reason, the problem of
providing and maintaining an impervious noble-metal barrier between the

*Based on material prepared by members of the Los Alamos Scientific Labora-
tories.
398 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [CHAP. 7

TaBLE 7-7
HRE-2 CoxstrucTioN Cost* (To MAY 1056)

Reactor (HRE-2):

Design
Reactor $549,000
Structure and services 185,000
Instrumentation 126,000
Total design cost (labor OH @ 659) $860,000
Reactor installation
Building modification 58,000
Reactor cell and shielding 375,000
Reactor components and piping 1,400,000
Reactor controls (instrumentation and valves) 470,000
Reactor steam system and cooling system 161,000
Miscellaneous service piping 69,000
Maintenance tools 54,000
Waste and off-gas system 78,000
Miscellaneous and spare parts 208,000
Division supervisory labor 120,000
Total installation (labor and material) $2,993,000
Total reactor cost $3,853,060

Containment cost estimate for HRE-2:

1. Additions to shield design, required only to meet containment

specification $75,000
2. Cost of sealing conduits and wiring 7,000
3. Cost of leak-tight closures on process service lines 23,000

4. Blast shields on reactor components

Pressure-vessel $35,000
Two steam generators 10,000 45,000
5. Radiation monitoring of service lines 10,000
Total cost $160,000

*Excluding $800,000 cost of supporting research and development labor.
tBuilding housed HRE-1—initial cost, $300,000.
7-5] THE LOS ALAMOS POWER REACTOR EXPERIMENTS 399

TaBLE 7-8
ProrerTiES OF LAPRFE-1 aAnp LAPRE-2 FurL SYSTEMS

 

 

 

 

 

 

 

LAPRE-1 LAPRE-2

Type of uranium oxide U0s U0,
Weight percent H3PPO4 53 95
Vapor pressure, psig 250°C 400 190
Vapor pressure, psig 350°C 1800 300
Vapor pressure, psig 450°C 4500 800
Overpressure gas 02 H,
U235 concentration, g/liter 111 77
H concentration, moles/liter (25°C) 90 62
H /U235 190 190
Fuel volume ratio (430°C to 25°C) 1.33 1.16
Temperature coefficient of

i reactivity (per °C) —8X 1074 —5X 1074

 

fuel solution and all structural metals with which it might come in contact is
extremely difficult to solve. This problem was not solved in the first ex-
perimental reactor to use phosphoric acid system, the LAPRE-1; however,
considerable improvements in fabrication and testing techniques have been
developed for use in the construction of LAPRE-2.

Of the wide range of possible combinations of uranium oxide, phosphoric
acid, and water, two systems appear to be of special interest. The first,
which consists of UOj3 dissolved in an aqueous solution containing between
30 and 60 w/o0 of phosphoric acid and pressurized with an oxygen over-
pressure, was used for the first Los Alamos Power Reactor Experiment
(LAPRE-1). The properties of this solution are given in Chapter 3 and
summarized in Table 7-8. The vapor pressure of this solution at the design
operating temperature of LAPRE-1 (430°C) is 3600 psi.

The second solution (to be used in LAPRE-2) consists of UO32 dissolved
m a 95 w/o phosphoric acid and pressurized with a hydrogen over-
pressure. As seen from Table 7-1, the vapor pressure of this solution at
450°C is only about 800 psi. Because of its reducing properties, the cor-
rosive behavior of this solution is such that anything below hydrogen in the
electromotive series is only slowly attacked. Since the attack rate is pro-
portional to the position of the metals in the series, possible materials
in decreasing order of usefulness are platinum, gold, carbon, silver, and
copper. Although neither silver nor copper is attacked at a rapid rate, both
metals undergo mass transfer, and suitable means for inhibiting this must
be found before they can be considered practical for lining the reactor.
400 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHaP. 7

Solution Steam Discharge Line
Storage
Emergency
—iL Dump Tank

        
   
  
 
  
  
  

Transformer

 

 

 

 

 

 

Feed Water Solution
Pump Transfer
Pump & Line

Exhaust
& Vent
Valve Control

 

 

 

Auxiliary
Equipment
Room

 

  
 
 
 
 

Control
Console

- Y : — And Racks
Reactor ! m J L
Vessel 1

Lead and Water Shield Walls

 

Fia. 7-23. Plan view of LAPRE-1 reactor area (courtesy of Los Alamos Scien-
tific Laboratories).

7-5.2 Description of LAPRE-1. The first experimental reactor,
LAPRE-1, was housed in a cell at the Los Alamos site which had been
built previously to handle highly radioactive materials. The cell was modi-
fied by supplementing the shielding with additional concrete on one side
and lead on the other sides and ceiling. A stainless-steel wall was also put
across the reactor end of the cell to permit filling with water for a neutron
shield. Equipment such as the solution transfer pump and sampling system
were located in the cell outside the water-filled portion, while other auxili-
aries were located outside the cell. These included a 12 gpm, 4000-psi feed-
water pump, three 6-in.-diameter, 32-ft-long underground solution meter-
ing and storage tanks, and an emergency dump tank. Valve controls were
located on the wall of the cell. Figure 7-23 shows a plan view of the
LAPRE-1 reactor area.

Figure 7-24 shows an artist’s sketch of LAPRE-1 in which the essential
features of the reactor are indicated. As will be seen from this sketch, the
15-in.-ID> eylindrical reactor contains a critical zone above the cooling
coils and a reservoir region below. This latter region, which contains a
hollow cylinder of boron to prevent eriticality, provided a storage place for
excess fuel and avoided the necessity for transferring the highly radioactive
fuel solution in and out of the reactor as its temperature changed. The
7-5] THE LOS ALAMOS POWER REACTOR EXPERIMENTS 401

4.~ Steam Qutlet

g Control

  
   
  
    
   

Baffle

Critical
Region

 
  

’
P MHeoters &
i Insulation

Heat
change:

Boron

Pump
" Impeiler

 

E ,
. o "

Fic. 7-24. Los Alamos Power Reactor Experiment No. 1 (courtesy of the Los
Alamos Scientific Laboratories).

large negative temperature coefficient of reactivity made it impossible to
compensate for the excess reactivity of the cold critical reactor and it was
not considered practical to change the fuel concentration as is done in the
HRIE-2. In the LAPRE-1, shim control was achieved by means of the
thermal expansion of the fuel. This was accomplished by initially filling
the reactor with the cold phosphoric acid fuel solution to a level about
7 to 8 in. above the cooling coils. At this level the reactor would become
critical on removal of the control rods, the solution would heat up, expand,
and gradually fill the eritical zone. On cooling, the reverse process took
place, making the reactor self-compensating. The reactor vessel was so
designed that at 430°C, the operating temperature, the fluid level would
reach a point Just above the flow baffle. In the actual experiment the
reactor was maintained at zero power while the temperature was raised
402 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

Water Steam
Manifold Mqifcid

    
 

 

Heat Exchanger
Graphite
# Reflecto

Platinum Draft
Tube {Critical
Region)

®._ BeO Source

.- Startup Source
Reflector Shim

- Fue! Transfer
Fuel ! tine
Storage § . i
Tank

Fic. 7-25. Los Alamos Power Reactor Experiment No. 2 (courtesy of the Los
Alamos Scientific Laboratories).

to 240°C by using electrical heaters located around the exterior of the
vessel. This was done to minimize the positive rcactivity change resulting
from starting up the fuel circulating pump, which caused a rapid injection
of lower temperature fuel solution from the reservoir into the core.

Heat removal from LAPRIE-1 was accomplished by circulating the fuel
solution at 600 gpm over the cooling coils by means of the centrifugal
sealed rotor pump shown in Fig. 7-24. The 12-gpm feed-water stream,
heated to 800°F in passing through the cooling coils, was simply discharged
to the atmosphere through a quick-closing throttling valve and steam
silencer as a means of dumping of heat. The possibility of fisston products
escaping through the steam discharge line was prevented by installing a
70-sec holdup line, suitably monitored for radioactivity, between the
reactor and throttling valve.

In case of accidental stoppage of the pump during full-power operation,
the reactor could be shut down with the control rods. Provision for dis-
charging the fuel solution by hand or through a rupture disk was included
for added safety.

The entire LAPRE-1 vessel and cell components exposed to fuel solution
were gold-plated, with the exception of the rod thimbles and draft funnel
which were made of, or coated with, platinum to minimize neutron ab-
7-5] THE LOS ALAMOS POWER REACTOR EXPERIMENTS 403

TaBLe 7-9

CHARACTERISTICS 0F Lo0S ALAMOS
PoweEr ReacTor EXPERIMENTS

 

 

 

 

 

 

!

LAPRE-1 LAPRE-2
Power level, Mw heat 20 0.8
U235 in core, kg 4.1 4.1
U235 total, kg 8.5 6.5
Fuel temperature, °C 430 430
Truel pressure, psig 4000 700
Steam temperature, °F 800 600
Steam pressure, psig 1800 600
Cost of components $250,000 $120,000

 

 

sorption. The 22 stainless steel cooling coils, 50 ft long, were clad with
6 mils of gold.
The characteristics of LAPRE-1 and -2 are summarized in Table 7-9.

7-5.3 Operation of LAPRE-1. Final tests on the LAPRE-1 system
were made with a 0.51 47 TO3 in 7.25 A H3PO4 fuel solution. Data were
obtained at room temperature in terms of control-rod position at delayed
critical versus volume of fuel injected into the system,* and results were
interpreted in terms of a simplified calculational model to obtain control
rod worths. Tor the five control rods, four located on a 31%-in. radius
and one central rod, measurements yielded a total worth of 6.39,. The
latter results were in good agreement with period measurements at cold
critical.  Also inferred from the data was an effective delayed neutron
fraction of 0.0001.

With the reactor filled to a predetermined level at room temperature,
initial heating of the system was achieved by means of electrical heaters
disposed around the outside of the vessel. Heat was applied until a core
temperature of 240°C was attained; heating by nuclear power was then
initiated. With nuclear power a core temperature of 340°C was achieved,
at which point the circulator was turned on. Only during the forced con-
vection phase of the operation was a uniform temperature distribution
established in the fuel solution; previonus to starting the circulator, the

*Detailed description of the nuclear data obtained during the test is set forth
in “Control Rod Worths vs. Temperature in LAPRE-1" by B. M. Carmichael
and M. E. Battat, TID-7532 (Pt. 1), p. 125, US.A.E.C., Technical Information
Service Lxtension, Oak Ridge, Tenn.
404 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION [cHAP. 7

averange core temperature wus higher than the corresponding reservoir
temperature.

Data were obtained during the course of the experiment in terms of
control rod positions versus core temperature at delayed critical. For the
operation without the circulator, thermocouples located at various points
in the system provided data from which average core and reservoir tem-
peratures could be estimated. With the circulator in use, the data for
control rod position versus temperature were used to infer the temperature
of the solution level at the top of the baffle. Above this point a further
increase in height did not contribute to the reactivity; hence the system
was solely governed by the temperature cocfficient of reactivity. Based
on the caleulated negative temperature coefficient for the system
(~8 X 10~ 1/°C") it was found that the rods were worth 30 to 40% more
at 390°C than at cold eritical.

A maximum temperature of 390°C at an operating power of about
150 kw was achieved in the experiment. No steam data were obtained
during the test because of a failure in the heat-exchanger assembly after a
number of hours of operation. Inspection of the assembly after a cooling-off
period indicated that a rupture in the gold cladding of two of the lead-in
tubes was responsible for the failure. This rupture was probably due to
a bonding between the gold cladding of the tubes and the gold plate of the
hallle, in conjunction with vibration of the tubes resulting from the opera-
tion of the circulator. Although no further tests with LAPRIS-1 were
conducted after October 1956, careful inspection of the reactor parts
after disassembly provided information valuable to the construction of
LAPRI-2.

7-5.4 Description of LAPRE-2. The second power reactor experiment,
LAPRE=2, was constructed in an underground steel tank 50 in. in di-
ameter and 20 ft long located at a site approximately 100 feet from
LAPRI-1. A sketch of the reactor is shown in Ilig. 7-25. Because of the
lower vapor pressure of the LAPRE-2 fuel solution the design and method
of controlling LAPRI-2 differ from LAPRE-1. As shown n the figure,
the reactor vessel has a simple autoclave shape with relatively thin
(5/8-inch) wulls, This compares with the 3-in.-thick vessel required for
LAPRE-1. The cooling coils in LAPRE-2 are located above the reactor
core, which has the advantage of providing a noneritical region for excess
liquid in case of overfilling. Cooling is accomplished by natural convee-
tion circulation of fuel solution over the cooling coils. Fuel circulation is
aided by an mverted platinum cone in the core region. A 1-ft-thick graphite
refiector surrounds the vessel. The inner 6 in. of this reflector can be moved
slowly to adjust the reactivity at varying fuel concentrations.

Cold eriticality is achieved by slowly filling the reactor with fuel solution
7-5] THE LOS ALAMOS POWER REACTOR EXPERIMENTS 105

from a water-cooled copper tank pressurized with hydrogen. The line
connecting the tank and reactor is gold clad (0.010 in.). After the reactor
becomes critical with cold fuel solution, further additions are made until
the desired operating temperature of 430°C is reached. A separate pancake-
shaped cooling coil, located just above the main tube bundle, serves as a
level indicator. A thermocouple in the discharge of this coil indicates
whether there is power removal and thus if the coil is immersed.

All parts of LAPRE-2 are completely clad with fifteenn mils of gold to
provide 1009, protection of the structural metal from the fuel solution.
This gold, as well as the platinum and noble metals used in LAPRIS-I,
are recoverable after conclusion of the reactor experiments.

The 600-psi, 600°F steam produced in the LAPRE-2 is of somewhat
lower quality in terms of power production than LAPRIE~1 because of
the lower operating pressures. However, at a design level of 1.3 Mw heat,
the reactor could produce 250 kw of electricity which is acceptable for a
remote power station.

The characteristics of LAPRE-2 are summarized in Table 7-9. Con-
~truction of the reactor proceeded through 1957 and was essentially com-
pleted i early 1958.
406 EXPERIMENTAL REACTOR DESIGN AND CONSTRUCTION  [cHAP. 7

REFERENCES

. A. M. Wrinserag, Whither Reactor Development, Inter-American Con-
ference, Brookhaven National Laboratory, May 1957, Nucleonics 15(8), 99
(August 1957).

2. C. P. Baker et al.,, Waier Botler, USAEC Report AECD-3063, Los Alamos
Scientific Laboratory, Sept. 4, 1944,

3. F. L. BentzeN et al., High-power Waler Botler, USAEC Report AECD-
3065, Los Alamos Scientific Laboratory, Sept. 19, 1945,

4. L. D. P. King, Design and Description of Water Boiler Reactors, in Pro-
ceedings of the International Conference on the Peaceful Uses of Atomic Energy,
Vol. 2. New York: United Nations, 1956. (p. 372)

5. C. K. Beck et al., Nuclear Reactor Project at North Carolina State College,
Nucleonies 7(6), 5-14 (1950); and Further Design Features of the Nuclear Reactor
at North Carolina State College, USAEC Report ALCU-1986, North Carolina
State College School of Engineering, January 1952.

6. R. N. LyoN, Preliminary Report on the 1953 Los Alamos Boiling Reactor
Experiments, USAEC Report CF-53-11-210, Oak Ridge National Laboratory,
Nov. 30, 1953.

7. J. R. Dietricu et al.,, Design and Operating Fxperience of a Prototype
Boiling Water Power Reactor, in Proceedings of the International Conference on
the Peaceful Uses of Atomic Energy, Vol. 3. New York: United Nations, 1956.
(p. 56)

8. L. D. P. King, personal communication, January 1952.

9. Reactor News, Nucleonics 15(2), R10 (February 1957).

10. S. E. BeavL and C. E. WiaTers, The Homogenecous Reactor Experiment—
A Chemical Engineering Pilot Plant, Chem. Eng. Progr. 50, 256-262 (May 1954).

11. C. E. Winters and A. M. WrINBERG, Homogeneous Reactor Experiment
Feasibility Report, USAEC Report ORNL-730, Oak Ridge National Laboratory,
July 6, 1950.

12. C. E. WintErs and A. M. WrINBERG, A Report on the Safety Aspects of
the Homogeneous Reactor Experiment, USAEC Report ORNL-731, Osak Ridge
National Laboratory, Aug. 29, 1950,

13. C. L. Secasgr and F. C. Zarr, HRE Design Manual, USAEC Report
TID-10082, Oak Ridge National Laboratory, Nov. 18, 1952.

14. S. Visner and P. N, Havnexreicu, HRE Experiment on Internal Recom-
bination of Gas with a Homogeneous Catalyst, in Nuclear Scr. Technol., USAEC
Report TID-2506(Del.), Oak Ridge National Laboratory, 1955. (pp. 73-90)

15. R. B. Bricas and J. A. SwartouT, Aqueous Homogeneous Power Reactors,
in Proceedings of the International Conference on the Peaceful Uses of Atomic
Energy, Vol. 3. New York: United Nations, 1956 (p. 175). S. E. BEarL and S.
Visner, Oak Ridge National Laboratory, 1955. Unpublished.

16. D. O. CampBELL, Decontamination of the Homogeneous Reactor Experi-
ment, USAEC Report ORNL-1839, Oak Ridge National Laboratory, June 1956.
17. C. E. Winrters, Oak Ridge National Laboratory, 1952, Unpublished.

18. 5. E. BeaLn, The Homogeneous Reactor Test, in HRP Civilian Power
Reactor Conference Held at Oak Ridge, March 21-22, 1956, USAEC Report T1D-
7524, Oak Ridge National Laboratory, March 1957. (pp. 11-47)
REFERENCES 407

19. 8. E. Beary, The Homogencous Reactor Test, in HRP Civilian Power
Reactor Conference Held at Oak Ridge National Laboratory, May 1-2, 1957,
USALEC Report TID-7540, Oak Ridge National Laboratory, July 1957.
(pp. 11-46)

20. 8. E. Beary, Homogeneous Reactor Experiment No. 2, 4th Annual
Conference of the Atomic Industrial Forum, October 1957.

21. 3. E. BeaLy, in Homogeneous Reactor Project Quarterly Progress Report
for the Period Ending Jan. 81, 1958, USAEC Report ORNL-2493, Oak Ridge
National Laboratory, 1958,

22. I. Seiewak and H. L. FarkexBerry, The Homogeneous Reactor Test
Mockup, in Proceedings of the Second Nuclear Engineering and Science Conference,
Vol. 2, Advances in Nuclear Engineering. New York: Pergamon Press, 1957.
(pp. 62-69)

23. D. S. Toowms, JR., Instrumentation and Controls for the HRT, Nucleonics
15(2), 48-52 (1957).

24, 8. E. Beawn and R. W. JurcenseEn, Direct Maintenance Practices for
the Homogeneous Reactor Test, Fourth Nuclear Engineering and Science Con-
ference, March 1958 (Paper 82); and USAEC Report CF-58-4-101, Oak Ridge
National Laboratory, April 1958.

25. T. RockwerL LI, Reactor Shielding Destgn Manual. New York: D. Van
Nostrand Co., Ine., 1956.

26. N, GrasstoxE, Principles of Nuclear Reactor Engineering. New York:
D. Van Nostrand Co., Inc., 1955.

27. C. D. Borp and O. Sisman, How To Calculate Gamma Radiation Induced
in Reactor Materials, Nucleonics 14(1), 46-50 (January 1956).

25, 5. GrassToNE, Principles of Nuclear Reactor Engineering. New York:
D. Van Nostrand Co., Inc., 1955.

20. 8. E. Beavr, Containment Problems wn Aqueous Homogeneous Reactor
Systems, USAEC Report ORNL-2091, Oak Ridge National Laboratory, Aug.
K, 1956,

30. P. R. KasTEeN, Operational Safety of the Homogeneous Reactor Test, USAEC
Report ORNL-2088, Oak Ridge National Laboratory, July 1956.

31. 8. E. Bearr and S. Visyer, Oak Ridge National Laboratory, 1955.
Unpublished.

32. S. E. BeaLL and S. Visner, Oak Ridge National Laboratory, 1955.
Unpublished.

33. P. M. Woon, A Study of the Possible Blast Effects from HRT Pressure
Vessel Rupture, USALEC Report CF-54-12-100, Oak Ridge National Laboratory,
Dec. 14, 1954.

34. D. Fromax et al., Los Alamos Power Reactor Experiments, in Proceedings
of the International Conference on the Peaceful Uses of Atomic Energy, Vol. 3.
New York: United Nations, 1956. (p. 283)

35. R. P. Hammonp and L. D. P. King, Los Alamos Homogeneous Reactor
Program, in [T RP Ciwilian Power Reactor Conference Held at Oak Ridge, March
21-22, 1556, USAEC Report TID-7524, Los Alamos Scientific Laboratory,
March 1957. (pp. 168-209)