3

CO
CO

THE STATUS OF SODIUM CLEANING

PROCESSES IN THE LMFBR PROGRAM, USA*

by
SfJ 04-
6. R. Taylor
E. C. Bishop
M. H. Cooper

Westinghouse Advanced Reactors Division

and
:
'sen*** • . - •••--j-u.

P. Perrone
U. S. Atomic Energy Commission, RDT
-NOTICE-
This report was prepared as an account of work
sponsored by the United States Government. Neither
the United States nor the United States Atomic Energy
Commission, nor any of their employees, nor any of
their contractors, subcontractors, or their employee],
makes any warranty, express or implied, or assumes any
legal liability or responsibility for the accuracy, com-
pleteness or usefulness of any information, apparatus,
product or process disclosed, or represents that its use
would not infringe privately owned rights.

*Work Funded Under AEC Contract AT(11-1)-3045, FFTF Pump Bearing Test

UIS7E.SUTI0H OF THIS BQCONIWT

 I. Introduction
The basic methods which have been used in sodium reactor programs to clean
and requalify specimens and components that have been exposed to sodium are
well known, although not well documented. Moreover, while these methods have
been extensively applied to laboratory cleaning of test specimens or small,
simple components, their use in cleaning complex reactor or test components
and systems, as well as requalifying them for continued service, has been
limited and sporadically unsuccessful. Documentation of the precise condi-
tions of reactants, temperature, time, etc., has been almost nil, with
the result that the definition of a reliable process and its embodiment
into a specification or standard for a complex component is not possible
at this time.

In view of the imminence of the need to clean and requalify large, intricate
components in LMFBR testing programs, as well as actual reactor components,
considerable engineering application work must be done to be prepared for
these requirements. It is the purpose of this presentation to recount the
results of several component and system cleaning cases, examine examples and
causes of successes and failures, and identify the further application
engineering they dictate should be carried out.

Table 1 gives a comparison of the advantages and disadvantages of the

principal cleaning processes. Inability or slowness of either alcohol or
vacuum distillation to remove sodium compounds, such as sodium oxide or
sodium carbonate, is a significant advantage for the moist nitrogen process.
Both the reactant cost, in the case of alcohol, and the engineering and
equipment cost for vacuum distillation are more serious disadvantages then
the formation of caustic solutions in the moist gas process. The catastrophic
effects of caustic stress corrosion cracking can be avoided in the moist nitrogen
process by careful temperature control. Sufficient experimental data are not
available now to judge the superiority of any process in cleaning crevices;
however, effective crevice cleaning has been reported for all three processes.
 TABLE COMPARISON OF SODIUM CLEANING PROCESSES

Process Advantages Disadvantages

Hoist Nitrogen 1. Reaction rate readily controlled 1. Formation of caustic solution which
by moisture to nitrogen ratio can lead to stress corrosion cracking
if temperature exceeds safe limit (/w250°F).
2. Moisture diffuses into entire 2. Reaction with sodium in crevices
volume, eliminating potential limited by diffusion of moisture through
unreacted areas because of gas crevice and absorption by caustic at the
trapping gas-liquid interface.
3. Reaction proceeds with both 3. Hydrogen is evolved.
sodium and sodium oxide
4. Reactants costs low
5. Reactants non-toxic

Alcohol 1. Reaction rate slow; can be 1. Does not react with sodium oxide;
controlled by water additions hence must prevent oxidation of sodium
during handling prior to cleaning.
2. Formation of solutions of 2. Alcohols flammable and toxic.
NaOH avoided
3. Low temperatures easily 3. High reactants costs may require
maintained distillation and recycle.
4. Reaction with sodium in 4. Sodium may not be reacted in areas
crevices limited by diffusion that form gas pockets.
of reaction products through
crevice to the bulk alcohol 5. Alkaline organic compounds formed,
which could cause stress corrosion
cracking if not renoved.
 -2-
Process Advantages Disadvantages
Alcohol 6. Hydrogen is evolved.
Vacuum Distillation 1. Removes sodium without forming 1. Non-volatile sodium compounds are not
a caustic solution distilled.
2. Cleaning of crevices possible 2. System must be caoabie of being heated
without leaving caustic residue uniformly at least to 800°F and being evacuated
to 100 microns or less.
3. Can be performed in situ for 3. In-ieakage of air can oxidize sodium
some cases using trace heaters and prevent distillation.
4. Reactant costs are eliminated 4. Scale up to large components complex
and waste disposal minimized and equipment is expensive.
5. Hydrogen '.. ~~* produced.
 to the component. Caustic stress corrosion cracking is
avoided by this cleaning procedure because the temperature
is low (i.e., steam temperature or below), caustic readily
drains from component, and exposure time is short.

2. Pumps
Subassemblies of the three Enrico Fermi-Power Plant primary
sodium pumps were successfully cleaned in 1964 after more
than 7500 hours operation by uncontrolled atmosphere steam
cleaning. Removal, disassembly, and cleaning of these
pumps were required to modify internal check valves to correct
a system hydraulic instability (water hammer). After withdrawal
Into an argon filled transport cask, the pump was allowed to
cool to ambient temperature. Then the pump was slowly lifted
from the transport cask into an air atmosphere and installed
in a disassembly fixture. As the pump was disassembled, the
subassemblies were taken outside for steam cleaning. After
inspection, Including dye penetrant testing of the Colmonoy
bearings, the pump was reassembled with the modified check
valve and re-installed in the plant. All three pumps were
processed using this procedure. Since the uncontrolled atmosphere
cleaning procedure was used for these pumps, there was no
damage because of the factors discussed previously.

3. Heat exchangers
ARD successfully cleaned an experimental sodium heated
steam generator using a modified uncontrolled atmosphere
cleaning process. The steam generator, which consisted of
three 1-1/2 inch diameter tubes, was cleaned outdoors by purging
each tube successively with argon and dry steam. Although
there were some small explosions and burning sodium ejected
from the tubes, no damage attributed to cleaning was noted by
visual and metallurgical examination of the sodium tubes. This
cleaning operation was successful because of its short duration,
low temperature, and unconfined volume for hydrogen reactions.
II. Component and System Cleaning Case Histories
A. Steam or Moist Inert Gas
As is well known, the hazard associated with "steam" cleaning
austenitic stainless steels is the potential of stress corrosion
cracking, either from chloride carryover in the steam or rinse
water, or from the caustic formed in the sodium/water reaction.
Chloride contamination can be controlled, but caustic is
inherent in the process.

Of the four parameters affecting caustic stress corrosion cracking

(CSCC)-caustic concentration, residual tensile stress, time, and
temperature - only time and temperature are controllable during
cleaning, since caustic over a wide concentration range and signi-
ficant residual stresses-are probably unavoidable. Figure 1
summarizes the data relating time and temperature to the onset of
CSCC for austenitic stainless steels. Note that data for times
less than two days is only semi-quantitative. Reference will be
made to this figure in the ensuing description of several steam
cleaning experiences.

The steam or moist inert gas method of sodium removal has probably
had the widest application over the past 15-20 years with many
successes, but some failures, covering a broad range of structures -
large and small, simple and complex. It's use with fuel assemblies,
such as SRE Core II' 1 ' and EBR I P 2 ' are well known. Remaining
applications may be grouped into several classes:

1. Uncontrolled Atmosphere Cleaning of Small Components

Uncontrolled atmosphere cleaning by flowing dry steam has been
successfully used In cleaning small components by many liquid-
metal workers. While small explosions, usually noted as bangs
and pops during steaming, result from reaction of evolved hydrogen
with air, damage to the components being cleaned are rare.
Presumably, the components are not damaged because the hydrogen
explosions occur In an unconfined volume, which Is often external
 Cleaning of the 10% Model of the S2G (Seawolf) Superheater
at KAPL was not as successful. Is)' Severe cracking of the
superheater tubing was observed during and after the
cleaning operation. The cleaning procedure specified the
use of superheated steam rather than water to reduce the
explosion hazard. However, the use of superheated steam,
presumably resulted in temperatures above the critical temperature
as illustrated in Figure 1 for caustic stress corrosion cracking
to occur at the exposure time of the system to the caustic.
4! Test loops
Steam or moist inert gas cleaning of a few test loops has
been reported. Although a 4" sodium loop at KAPL^ ' reached
temperatures as high as 1200°F during steam cleaning, no CSCC
occurred, presumably because cleaning times were short and no
aqueous phase was present at the extreme temperature. A Type
316 stainless steel loop of 75-100 gal. NaK capacity was drained
and successfully cleaned with an argon/steam mixture in the
CAMEL project^ 7'. No loop preheat was employed, but tempera-
tures were controlled to 200-300°F by gradually enriching
the mixture, while circulating it through the loop for about
one week. A water rinse and vacuum dryout at 100-200°F prepared
the loop for insertion in ETR, where it operated successfully
for 1000 hours. In this instance, although cleaning times
were long, CSCC was probably avoided by controlling temperatures
below 300?F.

A recent attempt to steam clean a sodium test facility, the

Pump Bearing Test Facility (PBTF) at the Liquid Metal Engineering
Center, resulted in extensive caustic stress corrosion cracking
of the loop piping. A P&I diagram of the sodium system, modified
for steam cleaning, is shown in Figure 2. The bulk of the piping
was 2" and 6" diameter AISI type 304 stainless steel, which had
been gravity drained prior to cleaning. However, significant
quantities of sodium remained in non-drainable areas, such
as blind spots, weirs created by valves (which were removed
prior to cleaning), etc. Loop piping was preheated, nominally
to 250°F, prior to passing a steam/nitrogen mixture, also
preheated to 250°F, through the piping. Piping preheat and super-
heated steam were utilized to increase reaction rates and,
mistakenly, to attempt the formation of anhydrous caustic,
which would not cause CSCC.

Although it was intended to clean the piping in sections, the

loop was actually cleaned as a system, and the steam/nitrogen
mixture and sodium reaction products were present in the
system for 10-12 days. Moreover, problems with system
preheat temperature control and the heat generated by sodium
water reactions resulted in actual system temperatures of
300-400°F in some regions for several days, and locally higher
temperatures for similar periods.

The results of this exposure are shown in Figures 3 and 4,

which illustrate the branching type of cracking typical of
transgranular stress corrosion. All cracks were associated
with welds, but were not located in the weld metal areas.
They were located in heat affected zones, presumably locations
of high residual stress. It is instructive to note that
cracks occurred on the inside diameter of 2" diameter x 0.154"
wall piping opposite O.D. thermocouple tack welds, even though
the fusion zone extended only 10-25% into the wall from the
O.D. The absence of cracks in the weld areas of the stainless
steel tank used to mix the steam and nitrogen, which was
exposed to similar time - temperature cycles, ruled out the
possibility of chloride stress corrosion from the chlorides
present in the boiler feed water. The valves and the drain
tank of the loop, which were separately steam cleaned at
temperatures of approximately 250°F for times ranging from a
few hours to slightly more than one day, revealed no cracking.
Much of the piping I.D. had a heavy, dark oxide coating,
often observed by others on stainless after exposure to
caustic at high temperatures. ....
Obviously, failures occurred because areas of the loop piping,
containing residual stresses from welding, were in contact
with caustic solutions for days at temperatures considerably
above the curve shown in Figure 1. Actually, analysis of the
data indicates that cracking occurred only.in areas that were
in excess of~300°F. The attempt to maintain anhydrous
caustic failed because, once formed, it will pick up moisture,
from even superheated steam and form an aqueous solution.

What are the lessons to be learned from these cases and

what are their implications for future cleaning of larger,
more complex components and systems? While it certainly can-
not be denied that a more precise definition of the time-
temperature boundary for the initiation of CSC3, especially
at short times, is required, the failures that have occurred
cannot be attributed primarily to the imprecision of this
data. Rather, this experience demonstrates the need to
concentrate cleaning effort on the engineering problems of
applying the known technology of this process to complex
structures and systems.

Unfortunately, although the application to many components

has beerr successful, little documentation exists to form a
basis of preparing a standard or even positive guidelines
covering the method. Moreover, the experience of PBTF
demonstrates that the lack of a clearly delineated set of
limits led to inadequate control of cleaning parameters. Once
the process had started, on-the-spot decisions were made to
permit deviations, simply because no strict limits were set.
The entire process was not thoroughly planned so that the
mandatory controls to avoid CSCC were identified and rigidly
maintained. The point'should be made that in cleaning
reactor components or systems, a similar result could occur
again unless more attention is paid to doing the engineering
application work required to truly adopt the process to such
components. The principal tasks involved in this effort,
such as mock-up or prototype cleaning, demonstrating requali-
fication, disassembly to verify cleaning and integrity, etc.,
are /Hcnmcprf latpv in t.hic
B. Alcohol
Alcohol has been utilized for cleaning large components because of
its slow and controllable reaction rate and lower risk of damage.
Despite its higher cost and potential fire hazard, alcohol has been
successfully used in cleaning large reactor components, such as
pumps and a steam generator. Careful attention to inert gas blanketing
was utilized in these cases to minimize the fire hazard. Components
to be cleaned by alcohol should be protected from air contamination,
since sodium oxide, hydroxide, and carbonate are non-reactive with
alcohol and are relatively insoluble.

Following the failure of the sodium steam generators at the

Enrico Fermi Power Plant in 1962, the steam generator tube
bundle was removed by withdrawal into a nitrogen filled shipping
container.'3' The container was transferred to an outdoor cleaning
facility equipped with a 50 ton crane, nitrogen supply, railroad
tank car with ethyl alcohol, and required piping. All vents were
fitted with flame arrestors and all components were grounded. The
tubes which had failed were plugged and alcohol pumped into the
remaining tubes at 20 gpm. Absence of hydrogen in the nitrogen purge
indicated that sodium had not penetrated any of the unplugged tubes.

The shell side was cleaned by pumping alcohol into the shipping
container at 10 gpm and sparging with nitrogen at 5 to 10 SCFM.
Filling was interrupted periodically to monitor temperatures and
pressures. During filling, the pressure increased about 3 psi
and the hydrogen concentration in the vent line varied from 20
to 65%. The maximum local temperature observed was about 100°F.
The tube bundle was then soaked in alcohol for a total of about
60 hours, with a partial drain and refill after 12 hours to
attempt to augment internal mixing. Then the alcohol was drained
and the tubes dried with nitrogen.

The cleaned tube bundle showed a small white deposit, identified

as sodium carbonate, above the sodium operating level * The sections
of the tubes immersed in sodium were covered with a fine gray
powder, reported to be Croloy metal particles resulting from
sodium reduction of the oxide film initially present on the tubes.

An EBR-II primary sodium pump was removed and cleaned after more than
32,000 hours of service because of shaft binding arising from oxidized
(9)
sodium deposits. The pump was withdrawn into an argon filled
caisson, allowed to cool, and the surface of the adherent sodium
oxidized by slow introduction of air. Then the pump was partially
disassembled and the parts transported to an outside metal-
covered, concrete cleaning pad. Pump subassemblies were placed in argon
filled tanks specifically designed for each unit. Ethyl alcohol
was remotely pumped into each tank until the tanks were filled. Water
was slowly added to the alcohol until a 50% mixture was obtained.
When the sodium reaction was complete, the alcohol was drained
and the subassembly washed with hot water. Subsequently, the pump
assemblies were decontaminated with Turco No. 4521 and rinsed.
Modifications to the pump to alleviate the sodium deposition on
the shaft were accomplished and the pump returned to service.
At the time of the referenced report, the pump had been operated
satisfactorily for more than 1500 hours.

Thus, two complex reactor components have been successfully

cleaned with alcohol, and one of them, the EBR-II pump^returned
to service. Well planned procedures and inert gas protection
during cleaning contributed to the satisfactory results achieved
in these two cases. The decision of the EBR-II staff to oxidize the
sodium surface is rather curious, in that sodium oxide inhibits the
reaction with alcohol. Presumably, the water added to the alcohol
reacted with this sodium oxide.

Both EBR-II pumps had been removed and cleaned in 1963 about six
months after start-up because of binding problems. Hence, at
least two cleaning cycles on one pump and a single cleaning on the
other has had no noticeable effect upon their performance.
C. Vacuum distillation
Vacuum distillation has been successfully employed on both a
laboratory scale and on small components or systems to remove
residual sodium. Emperically, its application to larger systems
appears attractive, i.e., temperatures are no higher than those
required for system operation and may be lower, pressures do.not
require elaborate pumping systems, and times may not be excessive.
However, even experience on a small scale and analysis of the
extrapolation to larger systems reveals that the engineering to
permit reliable use on large complex components is inadequate.

At the JW Advanced Reactors Division vacuum distillation has been

employed for cleaning a wide variety of components and systems,
with varying success:

1. A 37-pin, grid type dummy fuel assembly used in a sodium

fretting test was very successfully cleaned at ^ l O O C F for
316 hours at pressures of 20-200 microns. Even areas of fuel
pin-grid intersection and threaded end fittings (Figure 5)
were completely cleaned, with no trace of a residual "smudge"
often found after vacuum distillation.

2. In contrast, repeated cleaning of a sodium wear test rig at

/^800°F and 1-5 microns for 72 hours on each occasion has
resulted in the gradual accumulation of a crust on certain
areas, such as bellows. Since the rig operates with a considerable
cover gas volume, it is probable that this accumulation is
due to frost formation, containing oxide formed from oxygen
and water vapor impurities in the cover gas. Removal of this
residue is possible only be reaction to form a soluble reactant,
such as flushing with water.

Attempted vacuum distillation of a complex control rod sodium

test system (Figure 6) was only partially successful due to
 a variety of causes:

1. Frost or oxide formation in cover gas areas

2. Limitations on heating certain components above 400°F
3. Intricate, long-passages requiring long diffusion times
4. Inability to determine the completion of the distillation
process.
It was eventually cleaned using Dowariol; however* modifications
are being made to permit heating of the entire system to 800°F
and the use of gas sweeping to obtain more complete distillation
during the next cleaning cycle.

These and other similarly reported experiences presage the problems

of applying vacuum distillation to larger, more complex components:

1. Since oxides or other compounds of sodium are almost

certain to be found to some degree, especially in components
exposed to a cover gas, incomplete removal of sodium
will result.
2. Residual compounds left after vacuum distillation must
be reacted away, probably with water or alcohol rinses.
3. Since the advantages of vacuum distillation are enhanced
by performing it in situ (minimize oxidation, utilize
trace heating, isolate with existing valves, etc.) the
need to follow with a rinse presents special problems
in isolation from the remainder of the system. Vacuum
distillation does however simplify subsequent removal of
the component from the system for further cleaning.
4. Complex components which have long, intricate passages
create long diffusion times, unless gas sweeping is
used to accelerate vapor removal. Little experience
exists with such techniques, which also increase the
probability of oxidizing the remaining sodium.
5. With intricate components it is difficult to determine
the end point of the distillation process. Disassembly
 and examination Is likely, of course, to oxidize remaining
sodium.
6. All parts within an enclosed area to be vacuum distilled
must be capable of being heated to at least 800°F.
7. Heating of large, complex structures with trace heating In the
absence of sodium can result in thermal stresses, unless the
heating Is specifically designed for this purpose.
8. If the containment vessel for vacuum distillation Is a part
of the reactor system (pump tank, e t c ) it must be designed
for vacuum loads to temperatures of at least 800°F.
9. Sodium valves may not provide adequate isolation for vacuum
distillation and subsequent water rinsing.
III. Implications for Reactor Component and System Cleaning
A. Components
The experiences cited here, as well as those of others, serve
to point up the problems which can be anticipated when attempting
to apply these cleaning methods to larger* inert complex reactor
components* especially those ftm pritoary systeras, which will
exhibit radioactivity level* considerably In excess of any
encountered to date, it would appear that many of these problems
©f engineering application ef the known technology sf cleaning
have not been adequately addressed in component design, delineation
of actual process steps; and eitanina. facility design. Some of
Hie wore significant areas are discussed helm:

1* Disassembly
Both the successful £8ft»II and the Enrico Fermi pump cleaning
operations were accomplished with partial disassembly. This
disassembly simplified the cleaning operations and permitted
inspection of critical surfaces. The demonstration by
examination that cleaning had not damaged the ptimp contributed
greatly to the acceptance of the pump for reuse by the
reactor operators.

The capability to disassemble a component during cr after

cleaning is therefore clearly a desirable design objective.
Alternatively, If design for disassembly Is deemed impractical»
mock-up cleaning programs must demonstrate unequivoeably the
capability of the process to clean and maintain the structural
Integrity of critical surfaces in Identical configurations.
In designing for disassembly, the fact that primary system
components will be radioactive from transported corrosion and
fission products must of course be considered in specifying
procedures.
2. Isolation
Isolation can consist of removal of the component from the
system or isolation of the component in situ in the system.
The former is preferable* however, it may be impractical for
valves* pump and IHX tanks, as well as certain vessel internals.

Existing valves may not represent satisfactory means for

isolation. In general, sodium valves have a certain amount
of leakage and are not consistent with the isolation require-
ment. In tills event provision for capping off or inserting a
temporary block must be considered in the original system
design. In other words* the design must consider the alternatives
available for isolation of each component and either provide
that Isolation or identify a method for attaining it in the
event cleaning 1s necessary.

3. Orainability
Although drainability is generally identified as a design
criterion for sodium components, most complex components
are not designed for complete drainability of sodium as
illustrated by the PBTF loop, and cleaning processes must
be mated with the component to Identify drainability require-.
nents. It may be necessary to provide for methods of rotating
components during cleaning to present new surfaces to cleaning
agents and provide for draining others. Such handling provisions
must be made before operation of primary system components,
since activity levels may not permit it later.

4. Temperature Monitoring and Control

Since temperature control Is vital to most cleaning methods.
identification of the cleaning process prior to design is
highly desirable in order to incorporate the proper capability
 In the original design. Strategic location of thermocouples
to monitor the progress of the cleaning process at critical
areas should be a design requirement. Heating or cooling to,
maintain the desired temperature must also be provided in the
original design. Trace heating for sodium fill may not be
adequate because of capacity, control, or uniformity. In
the absence of sodium, the uniformity of heating is important
to avoid excessive thermal stresses. Where cooling is
required, analysis should verify that adequate heat removal
capacity is available during the cleaning process.

5. Design for Vacuum Distillation

In addition to heating requirements, if vacuum distillation
Is to be used for cleaning a component and its own contain-
ment Is to serve as the distillation vessel, it must be
mechanically designed to withstand a vacuum at 800°F
minimum. Attention should be paid in the design to the
location of evacuation system attachments and the possible
need to introduce a sweeping gas.

6. Requalification
Until recently little attention has been focused on setting
criteria for requalification and establishing the methods
for meeting those criteria. Requalification of most
components cleaned to date has consisted primarily of visual
examination and a few instances of return to service without
failure for a limited period. Restoration of surfaces to
their original state In most cases cannot be considered a
viable criterion of requalIfication, because sodium service
and cleaning methods do alter surfaces. However, specific
criteria for acceptability could be established for surfaces,
crevices, and certain configurations based on feature tests
and the application of selected non-destructive examination
methods. In this connection, the cleaning and destructive
examination of mock-ups or prototypes of major components
or their critical features, as discussed later, Is of major Importance.
8. Systems
The requirements described above for components are fully
applicable to systems* whether test or reactor. However, systems
present certain additional problems, discussed briefly in the
following sections. It should also be noted that certain complex
components should be considered* for the purposes of cleaning,
as similar to systems, for which then the following-sections
nay also be applicable.

1. Disassembly, Isolation and Draining

The experience in cleaning PBTF has demonstrated the
desirability of breaking down and isolating parts of
systems Into manageable components. For this purpose,
consideration should be given to placing tees or crosses
in strategic locations as a means of providing access for
plugging and cleaning sections of loop piping. Coupled
with true design for drainabiiity, these provisions could
guarantee that reaction products would not be In contact
with loop piping and components for long periods of time.

2. Temperature and Vacuum

Use of existing trace heating systems for temperature control
and monitoring should be a design requirement, particularly
if vacuum distillation 1s contemplated. However, grouping
of heaters per control system may not be satisfactory if
uniform heating and adequate monitoring cannot be provided.

3. Operation
Operating procedures should be geared to minimizing cleaning
problems. For example, draining a system following an oil
spill may produce a coating of oil over the entire system,
Instead of localizing it. Control of subsequent operating
temperatures following an oil spill could affect disposition
and/or decomposition of the oil following a spill.
 In any event, operating procedures should recognize that
removal of sodium to remove contamination should be avoided,
If at all possible. Alternate methods, such as decomposition
and trapping, should be planned prior to operation and be
Incorporated In operating and maintenance procedures.
Examples of successful disposition of oil spills by this
means are SRE* 1 0 ) and F E R M I ^ K

C. General
Basically, there Is an Industry-wide need for a set of clear,
unambiguous standards for moist gas, alcohol, and vacuum
cleaning, which, if adhered to. would assure safe, reliable
cleaning of components or systems. Before a final standard
can be written, however, additional work is required.

The basic thermodynamic and chemical data on the materials and

reactions involved in moist gas or alcohol cleaning are available
or can be estimated with satisfactory accuracy. What is needed
1s (1) data on the rates of the various reaction processes parti-
cularly on the rate of water vapor or liquid and alcohol reaction
with sodium In crevices and with sodium in bulk. This is
needed tc specify conditions which will avoid runaway reactions
1n the latter case, and which will assure complete reaction in
the former; (2) a better definition of general corrosion rates
and of the CSCC environmental bounds In aqueous alkali solutions.
While absolutely safe operating conditions can be specified now.
they may be too restrictive or difficult to maintain with absolute
certainty; accordingly, en effort should be made to broaden the
acceptable time-temperature conditions, if necessary by the
development of a modified process as, for example, the simultaneous
use of an acidic oxide gas (C02 or NO2) in concert with steam
to decrease the pH of the reaction products as the cleaning process
progresses; (3) once tentative standards are established, their
validity must be demonstrated by satisfactory application to the
 TABLE 2
CONTROLLING PARAMETERS OF LARGE SODIUM COMPONENTS
(Sheet 1 of 2)

How Maximum
Component Handled Weight(l) Length Height Diameter Cleaning Method

1. Medium Capacity Vertical 90 (113)tons<2> N/A<3> 40 ft ( 4 ) < 13 ft In cleaning vessel

Sodium Pump
Internal Assembly

2. Medium Capacity Vertical 65 (B2) tons N/A 40 ft 13 ft Out of cleaning vessel

Sodium Pump
Outer Case

3.a) FFTF Sodium Pump Vertical 64 [74] N/A 40 ft<4> < 6 ft In cleaning vessel
Internal Assembly

3.b) Pump Lift Fixture N/A 1.25 [1.4] tons N/A N/A N/A N/A
4. FFTF Sodium Pump Vertical 40 (50) tons N/A 34 ft 7 ft Out of cleaning vessel
Outer Case
5. AI Modular Steam Horizontal 14 tons 72 ft N/A 2-3 ft Out of cleaning vessel
Generator &. Vertical
6. FFTF Intermediate Horizontal 68 (85) tons 35 ft N/A 7 ft See Note (7).
Heat Exchanger
(complete)

7. FFTF Intermediate Horizontal 42 (53) tons < 35 ft N/A < 7 ft In cleaning vessel
Heat Exchanger & Vertical
(Tube bundle only)
(5)
8.a) FFTF Maintenance Horizontal 90 [iO4j N/A 59 •-9" 10 ft< 6 ) Requirement for cleaning
Cask (inc. hoist & Vertical not established
module)

8.b) Cask Vertical Lift N/A 2.8 [3.2J tons N/A N/A N/A N/A
Fixture

9. FFTF Dump Heat N/A 60 (75) tons • 35 ft 10 ft 16 ft wide Out of Cleaning Vessel
Exchanger
 cleaning of a series of complex prototypic systems and components;
(4) an answer to the question of whether or not even the best
sodium cleaning process which involves a thorough water rinse
as the final step does not detract from the acceptability of
the component for subsequent long term use in sodium. The plant
system components themselves must give the ultimate answer, but
the calculated risk can be considerably lessened by continuing
out-of-reactor studies.

D. Cleaning Facility Designs

Two major cleaning facilities are planned as part of the U.S.
LMFBR program; the Component Handling and Cleaning Facility
(CHCF) to be located at the Liquid Metal Engineering Center
(LMEC), and the Alkali Metal Cleaning Facility (AMCF), to be
located at the Hanford Engineering Development Laboratory (HEDL).

Both facilities in concept will use moist nitrogen as the

reference cleaning method but will retain sufficient flexibility
for later additions of alcohol or vacuum cleaning of large
components if necessary.

Component Handling and Cleaning Facility* '

The size of the components to be handled and cleaned -at CHCF is
illustrated by Table 2. This facility is to be housed in a
structure (galvanized steel siding) 60 feet wide by 80 feet deep
by 67 feet high. Considerable flexibility in handling and
disassembly is to be available; for example, the facility must
have the capability of removing tube bundles from heat exchangers
or steam generators in either the horizontal or vertical position;
it must be capable of rotating all components from a vertical to
horizontal position or vice versa both indoors and outdoors.

The building will, in addition to an internal bridge crane, either

be equipped with a roof mounted stiffleg derrick of about 100 ton
Future Applications
In order to focus clearly on the problems to be solved in liquid metal
component cleaning in the future, preliminary sodium cleaning procedures
for the FFTF IHX tube bundle and'the primary pump will be outlined for discussion.

A. FFTF IHX Tube Bundle

The size ( 81 inches O.D. by 32 ft. overall) of the'IHX tube
bundle can be appreciated from Figure 7 . Some idea of
its complexity can also be seen, but the myriad of crevices, semi-
stagnant areas, and fiat surfaces which may drain poorly can be
appreciated only after careful review of detailed drawings. The
impossibility of a complete post-cleaning inspection is obvious,
necessitating complete confidence in the efficacy of the cleaning
procedure used.

A complete description of the proposed cleaning operation is beyond

the scope of the paper; however, a procedure in outline might be
as follows (the reactor or system is assumed shutdown and cooled
to 400°F or less).

The primary sodium is drained by opening a valve in a 2" line off

the primary outlet line after isolating the affected loop by closing
isolation valves. The secondary loop is similarly drained, but this
does not drain the secondary side of the IHX. To remove sodium
from the secondary side, a complex procedure , some steps using
remote tooling, and others using temporary radiation shielding,
leads to the introduction of a dip pipe through the secondary inlet
downcomer to the bottom of the tube bundle. The sodium is then
removed from the secondary side by gas pressurizing the outlet
side. Clearly a several pound heel of sodium may remain after
this operation.

After this operation, the dip pipe is removed, and in largely remote
operations* pipe spool pieces are removed from the secondary inlet
capacity or have access to a mobile crane of like capacity.
Cleaning facilities will include a large vessel, 13 feet in
diameter by 48 feet high and an intermediate vessel 8 feet in
diameter and 16 feet high.

Alkali Metal Cleaning Facility^ 3 ) "

The Alkali Metal Cleaning Facility will provide for the removal
of nonradioactive residual sodium from large components and
equipment parts that have been exposed to a sodium environment
during nonradioactive LMFBR experimental activities.

The facility will be located in the existing High Temperature

Sodium Facility (HTSF), will consist of a vertical cylindrical
cleaning vessel, auxiliary equipment for the cleaning process,
waste effluent disposal systems, demineralized water syscem, and
necessary monitoring and control equipment for operation of the
cleaning process.

The vessel will be sized to accommodate large components, with

the In-Vessel Handling Machine (IVHM) and the Instrument Tree
(IT) controlling design. The vessel will have the capability
of adding spray nozzles and headers as required to insure sodium
removal from complex component configurations. Immersion to the
sodium wetted portion of the shield plug of the IVHM and IT will
be possible. The vessel will be suitable for service to 15 psig
and 275°F. Sufficient flexibility will be incorporated in process
control instrumentation and components to assure adaptability
to the standard moist gas cleaning process which is to be developed
and the system will be modular to facilitate subsequent removal
to FFTF.

It is to be emphasized that the CHCF and AMCF are in the conceptual

stage; as a result of the PBTF failure, the processes and methods
to be employed are to be reevaluated to ensure the avoidance of
caustic stress corrosion cracking.
 TABLE 2
CONTROLLING PARAMETERS OF LARGE SODIUM COMPONENTS
(Sheet 2 of 2)

How Maximum
Component Handled Weight*1' Length Height Diameter Cleaning Method

10. FFTF Hot Leg Horizontal 9,356 lb 96 in. 8«9«W 54 in. In cleaning vessel
Isolation Valve

11. FFTF Cold Leg Horizontal 6,800 lb 78 in. 7i6«(8) 40 in. In cleaning vessel
Isolation Valve

12. FFTF Cold Leg Horizontal 1,340 lb 58 in. N/A 32 in. In cleaning vessel
Check Valve

13. SPTF Venturi N/A 1,500 lb 8 ft N/A 18 in. In cleaning vessel

Flowmeter

14. SPTF Valves:

a. 12 in. Isolation N/A 900 lb 32 in. 100 in. 26 in. In cleaning vessel
depth
b. 18 in. Control N/A 2,200 lb 38 in. 80 in. 40 in. In cleaning vessel
depth
ROTES:
(1) Figures in this column, unless otherwise noted, do not include weights of lifting or handling equipment or steel
shot shielding where applicable. ( ) represents actual weight plus a 257. contingency, [ ] represents actual
weight plus a 157. contingency.
(2) The handling equipment for the Medium Capacity Sodium Pump is estimated to weigh 12 tons (107. of pump internal
assembly weight including contingency allowance).
(3) N/A = not applicable.
(4) This dimension includes the shaft projection above the mounting plate and bearing housing.
(5) Maximum CHCF design load is 182.6 tons which consists of components Nos. 3 and 8. The 182.6 tons includes a
157. contingency.
(6) The cask valve is < 11 ft square.
(7) If the IHX is to be cleaned while fully assembled, the out-of-vessel method will be used.
(8) This is the height dimension to the top of the operator/fresze seal mechanism.
 and outlet lines, the secondary system is sealed, the shear
blocks removed, the omega seal cut in two places, the maintenance
cask sealed over the IHX, the IHX tube bundle pulled up into
the cask, a dummy tube sheet seal installed on the secondary system,
and the maintenance cask with the supported tube bundle contained
in an inert atmosphere is ultimately transferred to a railroad car
for transport to the sodium cleaning facility. (If the operation
is necessary before a maintenance cask is available, isolation
will be obtained by plastic bagging using techniques available or
to be developed at LMEC for the IVHM.) Here, the tube bundle will
be transferred to an inerted cleaning vessel, which, could most
conveniently be the spare prototype IHX shell. The shell would
be permanently installed with a 5-10,000 gpm low head pump,
storage tank, and installed piping and available piping sections
making it possible to pipe the IHX primary and secondary sides in
series as illustrated in Figure 8. This arrangement is required
to assure sufficient circulation of fluids in the final cleaning
stages to complete reaction in crevice areas and to accelerate
final cleanup of the rinse water by ion-exchange.

If alcohol cleaning is used, the entire circulating system

including the IHX bundle is maintained under an inert atmosphere,
and Dowanol or other slow acting alcohol is slowly fed-into the
system. Gas venting and alcohol addition are adjusted as required
to maintain safe conditions. When the system is full, circulation
is started and maintained until reaction is complete, water being
added if desired or mandated to speed up or complete the reactions.
When reaction is complete, the alcohol solution is dumped to recovery
or disposal system, and the system is flushed with pure water,
ultimately using circulation through a demineralizer system to
establish required purity for the draining and drying operation.
If radioactive decontamination is required, it will precede the
final purification step. Presumably after these operations, leak
detection and tube plugging operations can be carried out by
essentially contact procedures. Also, removal of the bundle from
the tank and visual inspection can be undertaken.
 In carrying out these operations* all of the principles and
guidelines discussed previously and yet to be established will
of course be applied.

Bv. FFTF Pump

The FFTF primary pump assembly, shown in Figure J3 , is about
40 feet high from the primary sodium outlet to the coupling
spool. On removal from the tank, the maximum diameter (at the upper
flange) is about 7-2/3 feet. As with the IHX, the figure
illustrates size and gives some idea of the complexity of the
component. The lower 12 feet of the removed assembly can be
detached by unbolting, leaving the exposed impeller with the
upper portion.

Removal of the pump from the tank and transfer to the cleaning
and maintenance area in the maintenance cask or bagging requires a
series of operations closely paralleling those used for the IHX
tube bundle. One of the design requirements of this pump 1s
that draining the primary loop should drain the pump, which has
sodium up to within a few inches of the temperature radiation
baffles during operation. The criterion of drainability
requires the inclusion of many slots and small by-pass flow paths
which are otherwise unnecessary or undesirable; accordingly,
draining may be slow and perhaps incomplete. It is nevertheless
proposed that the first step In cleaning the pump be the use
of moist inert gas. To do this will require a tank about 38 feet
high and 8 feet in diameter flanged to accept the pump flange. Use
of a prototype or spare pump tank, paralleling the approach used with the
IHX is not recommended since it would not result in significant flow
in the upper portions of the assembly.

A final decision has not yet been reached on at what stage in

the cleaning procedure the inlet-outiet-diffuser section should
be removed from the upper portion of the pump. Based on the
assumption that the pump will be radioactive, the feeling at
 References:

1. NAA-SI-388

2. EBR-2 Fuel Cleaning H FEF/S - Procedure FCF-OM-281.4, Revised April

1971, "Standard Subassembly Transfer and'Sodium Cleaning Operation."

3. T. P. Roso, "Primary Sodium System Pumps in tne Enrico Fermi Atomic

Power Plant," APDA-3O9 (Atomic Power Development Associates, Inc.},
June 1969.

4. G. R. Taylor, ARO, persona} communication.

5. 3. L. Van Uiien and R. 0. Wylie, "Metallurgical Problems of Liquid

Metal Heat Exchangers," Unnumbered Report, Knolls Atomic Power
Laboratory, January 14, 1957.

6. KAPL-M-WWK-4, "Cleaning of 4-inch Sodium Loop at Alplans Base by Two

Methods, Alcohol and Steam, May 14, 1954.

7. R. Cochran, personal communication.

8. R. H. Anderson, J. A. Ford, P. S. Lindsey, "Design, Fabrication, and

Preliminary Operation of the Steam Generators for the Enrico Fermi
Atomic Power Plant, APDA-307 (Atomic Power Development Associates, Inc.),
August 1968.

9. B. C. Cerutti, G. R. Deegan, J. D. Nuiton, W. H. Perry, and R. E. Seever,

"Removal and Repair of EBR-II Primary Sodium Pump No. 1," ANL-7835
(Argonne National Laboratory), May 1972.

10. "SRE Operates Again with Core 2..." Nucleonics, Vol. 20, No. 1,
January 1962.

11. W. L. Chase, E. C. Kovacic, et al, APDA Technical Memo No. 35, "Carbon
in the Primary Sodium System of the Enrico Fermi Atomic Power Plant,"
August 1, 1963.
present is that the cleaning should be carried through to a
full water rinse in the inerted tank, before any disassembly,
and hence, even in the non-radioactive condition the same
techniques should be used. After this, the pump can be removed
for disassembly in air, using semi-remote tooling and/or localized
shielding. Once the lower assembly and impeller are removed, the
radioactivity of the upper portion will probably be low enough to
permit essentially contact inspection and maintenance operations,
including further cleaning if necessary. The lower assembly and
impeller can then be radioactively decontaminated in intermediate
or small-sized vessels if required.

Clearly, for both the IHX tube bundle and the pump assembly, the
specification of time-temperature-concentration conditions and the
establishment of decision points in the procedures awaits the output
of the development programs.

Similarly, detailed design of some aspects of the necessary cleaning

facilities must await development results, although there can be
little doubt that a large, versatile facility will be required at
HEDL as part of the FFTF complex. - —
References (cont'd.)

12. HN 8112.1, Conceptual System Design Description, CHCF, Volume 1,

August 1971.

13. Project WCP-7O4, System Design Description for the Alkali Metal
Cleaning Facility, December 1972.
 700

600
ESTIMATED CLEANING EXPERIENCE: WITH NO FAILURES
EBR H FUEL ASSEMBLIES
500 SRE CORE H FUEL ASSEMBLIES
WESTINGHOUSE AL BONDED STEAM GENERATOR
SNR PUMP
BR-5 PBTF FAILURE
FERMI PUMP
S- WO - \ KAPL LOOP
•

CRACKING
£5 300 - \

STAEHLE
200 REF-COO-2018-21* PBTF
NO FAILURES

100 NO CRACKING

6 7 8 9 10 II 12 13
TIME (days)

Figure 1. Reported Caustic Stress Corrosion Behavior in 300 Series Stainless Steels
 bo
a
•H

0)

u
o

5609-2
 rl

(THESE CRACKS EXTEND THROUGH THE WALLS OF.

•i BOTH THE PIPE AND THE ELBOW

Figure 3. Results of Dye Penetrant Examination Showing Cracks Associated

With Tack Welds and a Butt Weld on PBTF Piping

Figure 4. Typical Transgranular ID Cracks in 2 in. Pipe From PBTF

Mag 100X
5609-3-4
 ™ * *-•* ^ . < r -i«» -*•<. v i ^ - •*• • ^ JJJ^ -..^^ . — «•• —'.

3 £ ^ - ; c ^ •*--* -.-r-*r r:.: -^*--^*^ •

i
I;
ji
T »'
li
;
1 i''-^. E *
?J
'"< ii'*-
S i ••• W
4 i-v 0.
7v-

• -" -ft*.-'

Figure 5. Top of SAFT Fuel Pin Bundle Showing Cleanliness After Vacuum
Distillation Cleaning
5609-5
 / / / / / / /

>^— LEVEL VESSEL /

CONNECTION ( l i q u i d )

if
— ELEV -227.00 I!

Figure 6. Part of Control Ro(| Assembly Illustrating Complex Configuration for Cleaning
 SOOMMIT K I T

SECOMMXT OUTUT

SB1
UPPER TUBE SHffr

PMUIYMICT
HAT TMHSfa TUBES

M E BMHSKW BENDS BTMSS SEAL

SUMO VESSEL

PWWHOBTUT

Figure 7. Intermediate Heat Exchanger and Guard Vessel

5609-7
 GAS VENT AND ANALYZER

,t
JIL_ INERT GAS PURGE
I I-
I I
f I
STORAGE
SURGE
TANK

CHEMICAL
SOLUTION
SUPPLY

ION
EXCHANGE
COLUMNS

y SOLID LINES ARE PERMANENT

CLEANING FACILITY LINES

DOTTED LINES ARE INSTALLED

FOR IHX TUBE BUNDLE CLEANING

Figure 8. Schematic of FFTF Tube Bundle Installed for Alcohol Cleaning

5609-8
 MOTOR SUPPORT STAND COUPLING SPOOL

OIL SEAL-BEARING ASSEMBLY

TEMPERATURE
RADIATION BAFFLES

TANK

GUARD VESSEL

BEARING

OIFFUSER VANE

FLOW STRAIGHTENING V A N E ^ I

6UARD VESSEL SUPPORT

Figure 9. Primary Pump With Guard Vessel

5609-9