s 2T T
0 r"l 1'% A

S lu.'.l'-' ~ 3

iy EETES
E A 'J'-I-H'": ".'.
iR § wee®

OAK RIDGE NATIONAL LABORATORY

gy P
e W

OFERATED BY For Internal Use Only

 

UNION CARBIDE CORPORATION
NUCLEAR DIVISION

2

ORNL <Y

PO S 15 O CENTRAL FILES NUMBER

e SEPO4 WK CF-70-9-3

 

 

DATE: September 1, 1970 copyng, 136

SUBJECT: Critique of the Molten-S5alt Reactor Experiment:
A Collection of Comments Submitted by
Persons Associated with the Reactor

TO: Distribution

FROM: R. H. Guymon

ABSTRACT

The MSRE was shut down in December, 1969 after over L years of suc-

cessful operation. As an aid in profiting from the lessons of

the MSRE,

critical comments were solicited from project personnel. There were

28 replies, which are reproduced and summarized in this report,

touching

on practically every phase of the MSRE. Many of the comments are

applicable to other reactors or similar projects that might be
taken.

Keywords: MSRE, reactors, fused salts, critique, design,

under-

development,

construction, training, operation, maintenance, remote maintenance,

personnel, management, communication, documentation, computers.

NOTICE This document contains information of a preliminary nature
and was prepared primarily for internal use at the Oak Ridge
National Laboratory. It is subject to revision or correction and
therefore does not represent a final report. The information is
only for official use and no release to the public shall be made
without the approval of the Legal and Information Contrel Depart-
ment of Union Carbide Corporation, Nuclear Division.

 
CONTENTS
Introduction « « =« « « « « «
Summary of Contents-
Comments as Received * * = = * =+

Appendix I

A Request for Criticism of the MSRE |

Appendix IT

List of Contributors_

Appendix III
Topical Index . . . . . .

16

82

83

8l

 
INTRODUCTION

In Janusry 1970, shortly after the MSRE was shut down, a letter was
sent to T6 engineers, chemists, and administrators who had been associated
with the reactor. The letter (a copy of which forms Appendix I) announced
the intention to publish a critique of the MSRE and solicited contribu-
tions. Although contributions were received from less than half of those
contacted, there were many good comments and suggestions which should be
of value to future projects.

In the interests of transmitting their messages unchanged, each of
the 28 replies is reproduced without any editing in one section of this
report. To alid the reader in finding comments on areas of particular
interest, in this section I have indicated in the margin the topic being
discussed in the text. In addition, a topical index is provided as
Appendix III.

As a further aid to the reader I have combined and summarized every-
one's comments. At the end of each paragraph of the summary, I note the
reply and page number where the ideas summarized sare originally expressed.
There were, of course, many duplications among the independent replies,
as may be seen from the multiple references at the ends of some paragraphs.

Additional specific comments and recommendations will be published

in reports on MSRE systems and components1 and on training and operation.2

 

1
R. H. Guymon, Editor, MSRE System and Components Performance,

USAEC Report ORNL-TM-3039, Oak Ridge National Laboratory, in preparation,

’R. H. Guymon, MSRE Operator Training and Operating Techniques,
USAEC Report, ORNL-TM-3041l, Oak Ridge Natiocnal Laboratory, in preparation.

 
SUMMARY OF COMMENTS
General

It is hoped that those responsible for various aspects of the molten
salt reactor effort will not only study this critique but also reread the
relevant portions of the progress reports and summary reports and make
the next reactor even better. (M-L9)

The success of the MSRE was enhanced by the ability and dedication
of the perscnnel associated with the project. Management showed their
interest and helped emphasize the importance of the job. (M=-53)(T-68)

Administrative and technical responsibility should be well defined.
Authority should accompany the responsibility. (s-6k)

Cooperation between all groups is a necessary part of successful
ocperation. Their separate interests should be encouraged. For example,
the analysis group should propose any experiment which might supply use-
ful information, the chemists should request samples of any type, and
maintenance should propose short cuts in repairs or modifications. Opera-
tions should review each thoroughly and should be encouraged to be conserva-
tive in order to maintain safe conditions for the protection of personnel
and equipment. (D-25)(M-51)(M-52)

An effort should be made to assure that all involved personnel feel
that they are on the reactor team. This includes the shop foremen, crafts-
men, and remote maintenance personnel. They should be kept informed as to
what is being done and why it is important. (A-16)(A-1T)(T-65)(T-66)

An effort should be made to retain as many of the original personnel
as possible. This includes supervision, operators, foremen, and crafts.
Operators should be on-site early to become familiar with the system and
to follow construction to assure that all systems will operate properly
and to label all equipment when installed. (K- U4l4)(K-L5)(K-LT)

Criteria should be established early and all key personnel should be
thoroughly familiar with them. Limits should be set at reasonable values
with changes made as infrequently as possible. (D-23)(H- 34) (K- 43)(U- 70)
(U-T1)

On a reactor experiment, maximizing reliability should take second
Place to making detalled analysis of the system performance. This applies
to such things as design, taking samples, and varying operating conditions.

(B- 20) (M- 51)

 
Sufficient instrumentation should be provided so that: (1) detailed
enalysis of the system performance can be made, and (2) the operator does
not have so many jobs that he can not handle them properly. (D-23)(E-26)
(E-27)

There should be considerable improvements in instrumentation and com-
puters on the next reactor due to advancements in the state-of-the-art.

(Q-62) (Q-63) (U-70) (U-T73)

Reactor equipment should be robust and fooclproof and when possible
the design should be such that operation can be done by rotating shift
personnel. Areas requiring more intimate relations between operator and
equipment should be identified early and special operating perscnnel
should be provided. Equipment whose operation is something of an art
should be minimized. (D-25)(I-37)(K-4L)(M-51)(0-5T)

During design, construction, and preparation for operation, all
phases of the operation should be considered. Safety, containment,
shielding, etc., should be adequate during power operations, shutdowns,
and various transient conditions. Advanced plans should be made for de-
commissioning the reactor. (K-U44)(K-L45)(K-L48)

Quality assurance should be given a prominent role in construction
as well as subsequent maintenance and modification. Adequate persocnnel
should be available. Samples of all materials used in construction should
be retained. (E-29)(M-52)(N-55)(T-65)(T-68)

In selecting a supplier, the proper weighting should be given to the
various aspects of the Job so that he will at least be an expert in the
most important phases. (J-38)

Savings of time and money (such as the decision to operate with mar-
ginal amount of flush salt, operating with an oil leak in the fuel pump, or
changing the specification for the fuel-pump motor pressure vessel from
stainless to carbon steel) should be carefully considered. Some prove
costly in the long run. (A-16)(H=-35)(I~36){J-39){(M=54)(U-T0)(W-T6)

When items are used repeatedly such as the sampler transport tubes,
the cost of decontamination should be weighed against the use of low-cost
throwaway items. (I-36)

When repetitive jobs are done (such as removal of graphite specimens
and containment startup check lists) extensive planning should be done to
reduce downtime. (K-L43)(1- L8

Behavior which may appear as a small sideline effect in a development
test should be carefully evaluated for it may become a major problem in
reactor operation. One should not rely on makeshift procedures but
should develop a solution which will work in the reactor. (M-50)
Procedures should be written in advance for checking out instruments
after installation or repairs and for doing maintenance. (K-LL)(K-Lg)

Design

All design calculations should be formally reviewed and reliable data
on the physical properties involved should be available. Thermal stresses
should be carefully considered. (H-34){L-L8)(M-51)(AA-T8)

Detailed design should not be started too early. Title-I design
should be completed before starting Title II and elementary-type drawings
should be issued and approved before detailed drawings are started. (K-43)
(U-70) (Y=-7T)

The use of conventional equipment should be encouraged, however, the
specific application should be carefully evaluated to assure compati-
bility. (H-35)

The design should have some flexibility in the form of spare con-
nections, redundant lines, and open space for additions. With sufficient
redundance and flexibility, it is often possible to circumvent unantici-
pated difficulties. {(M-52)(M-54)(V-75)

The ambient conditions in which equipment and instruments must operate
should be considered in the design. (L-L48)

Absclute instead of gage pressure transmitters should be used where
secondary containment is needed. (U-T73)

Thermocouple wire and other low-level signal wiring should be twisted
and shielded. (U-T3)

The use of a computer for keeping track of control circuit inter-
connecting wiring should be considered. (U-TH4)

When continued operation is critical, spare equipment powered by
separate power sources is recommended. (F-30)

The design of all equipment should be above the shut—off pressure of
the supply (for instance, the design pressure of the thermal shield was
lower than the shut-off pressure of the treated-water pumps). (L-L49)

Keep as much equipment as possible outside of the reactor cell. (Z-78)
To minimize thermal stresses and reduce operator effort, automatic

heater control should be provided. More separate controllers were needed
to get a more even temperature distribution. (D-2L4)(F-33)(N-55)

 
Documentation

Certain key documents such as the design report, flowsheets, ele-
mentary control circuit drawings, and electricel elementary drawings
should be issued as soon as possible and should te kept up to date
throughout the life of the reactor. (K-39)(K-L0)

Adequate drawings should be issued to supply the needs of all phases
of the project. For instance, construction pecple will be interested in
the pedigree of all material used, types of welds, etc. Exact locations
are often not too important. Remote maintenance personnel are not inter-
ested in the type weld but want to know the exact as-built location of
everything. Overly dimensioned isometrics and photographs would probably
serve their needs best. Both of these can usually afford to spend time
searching for information. On the other hand, operating people need to
know what thermocouples and other instruments are installed, where the
pipes go, etc., and they often need to be able to find this information
quickly. (A-16)(A-18)(A-19)(I-36)(K-L1)

A consistent integrated scheme of documentation should be established.

(K-L42)

Recording of too much information may be cumbersome but it may pre-
vent permanent loss of important data. Errors or mistakes should also be
recorded so that others can profit by them. Field changes should be well
documented. (K-L40)(K-L46) (K-LT)

Training and Operation

 

In preparation for operation it is important to maintain a list of
all variables with their normal values and limits. The appropriate project-
associated personnel should supply the limits and update these as required.,
This then becomes the basis for the logs and is helpful in writing the
operating procedures. Operating procedures should be written as early as
possible. (XK-LkL)(K-L5)

An operator training manual should be written for use in initial
training and the training of replacement personnel. More use should be
made of simulators and programmed instruction. Graphic panels, the jumper
board, and the model were helpful on the job training aids. Periodic re-
views and tests are recommended. (D-23)(K-L45)(0-56)(K-LT)

The training instructors should be trained in teaching methods. Super-
visors should be trained in the art of supervision. (XK-L46)(X-7T)

Up-to-date operating aids such as flowsheets placed in the various
operating areas were helpful as was having limits marked on heater con-
trollers. (0-56)(0-58)

 
10

Maintenance

Computerized programmed maintenance of equipment should be carefully
set up and periodically reviewed. (K-47)(V-T7L) -

Remote maintenance supervision and crafts need training early in the
program. (A-16)(A-1T)

Although all required remote maintenance jobs were successfully com-
pleted, hindsight indicates that provisions for this work could have been
better. (A-16)

Provisions should be made for removing equipment from the reactor
cell to a sealed shielded work area without going through the high bay.
Adequate room is needed for doing remote maintenance and areas involved
shou%d be well-ventilated and easily decontaminated. (A-1T)(V-T4)(V-76)
(V-T6)

Since the remote maintenance shield is a key item during shutdowns,
it should be of better quality. A sufficient number of shielded carriers
of all sizes should be available. A contaminated work area should be
provided. (A-17)(V-75)(V-76)(Z2-78)

The number of bolt head sizes used in the cells should be kept at a
minimum and the quality should be improved. (A-18)(V-T5)

Sampling and Analysis

 

On-site analysis of samples is desirable. As many as possible should
be in-line devices. Flow-through samples are more representative than
dip samples. ©Some specific analytical methods need improving such as
U*/U" and carbon content of the salt and tritium in various gaseous
effluents. (B-20)(B-21)(c-21)(K-47)

Cross contamination of samples should be carefully considered and
sampling mechanisms should be provided at other locations such as the
drain tanks. Efforts should be made to decide ahead of design or con-
struction what type samples will be needed and from where. (B-og)(B-20)
(B-21) (E-28) (I-36) (BB-80)

More consideration should be given to the chemistry of auxiliary
systems. When methods are not available to test for deterioration (such
as the lube o0il for the circulating pumps) periodic replacement should
be made. (F-30)(K-4T)
11

Computer

All reactors should have an on-line computer as at the MSRE. The
reactor design should be compatible with the computer and planning and
programming should start early. (E-27)(M-54)(U-T1)

Reactor engineers should be more closely associated with the computer
and should be trained so that they can participate in the programming of
it and fully utilize its capabilities. ORNL maintenance people should be
trained early. (K-46)(0-57)(Q-60)(Q-61)(Q-62)(Q-63)

During installation and checkout and subsequent maintenance or
trouble-shooting, the computer should be isolated from operating personnel
to avoid loss of confidence in the machine. (Q-61)(Q-61)

On future reactors, some routine control should be delegated to the
computer. The alarm functions should be improved so that important alarms
are not missed or ignored. (D-25)(0-56)(U-71)(U-73)

Noise analysis of all rotating equipment as well as neutron and
pressure noise analysis should be computerized. Proper instrumentation
should be designed for accomplishing this. (E-27)(G-33)

Adequate hardware diagnostic programs should be required of the com-
puter manufacturer. (Q-63)

Chemical Processing

The development and redesign of the fluorine disposal system should
have been done earlier and preferably in another facility. More develop-
ment work is still needed on the use of KOH-KI for removing excess
fluorine. (M-52)(P-59)

Fluorinators should have frozen salt walls or use an alloy not con-
taining molybdenum. (P-59)

All process gas lines leaving shielded areas should contain fibrous
metal filters for remcving metallic fission preducts. (P-60)
12

Miscellaneous

Don't use an existing building. (X-y7)(U-70)(Z-78)

Straight rigid control rods would be simpler and more reliable.
These should be used and placed where they would do the most overall good
even if extra vessel penetrations are necessary. (E-28)

Possible plugging of gas lines should be considered. Better valves
should be developed. The use of more filters is suggested and dual or
larger lines with installed heaters are recommended. (E-29)(F-32)(I-37)
(0-58) (W-76) (U-73) (BB-80)

Salt line penetrations through cell walls require careful design to
prevent plugging. {P-60)

More use should be made of two-out-of-three coincidence leogic to
permit testing and there should be more separation between safety and
control. (U=-T2)

Process radiation and other instruments should be calibrated in terms
which are physically significant. (E-28)

Leak-detected flanges should be replaced with welded pipe using re-
mote welding techniques where possible. Electrical connections in the
cell should be welded to prevent them from coming loose. (F-30)(N-55)

Methods of improvement on the use of freeze valves would be to ele-
vate or pressurize the drain tanks. Two FV's could be used in series with
alternate thawing to increase the life expectancy. Freeze valves made by
spinning INOR-8 pipe might be an improvement. (F-31) (F-32)

It is suggested that a low-speed large diameter pump integral with
the reactor be considered and that cerro alloys be investigated as a re-
placement for the coolant salt. (F-32)

It is suggested that tests be run with the system at negative pressure
to test the effect on fission product stripping. (F-32)

A number of the comments made in the HRE and HRT critiques,,aaL+ are
still applicable. (V-T7L)

 

3¢. E. Winters, Criticism of the Present Design of the HRE, ORNL-
CF-52-11-139, unpublished internal document, November T, 1952.

*s. E. Beall, Critique of the Homogeneous Reactor Test (HRT) and the
Chemical Plant (HRT-CP), ORNL-CF-61-6-116, unpublished internal document ,
June 1, 1961,

 

 
well

13

The holdup time of the charcoal beds appeared to be only about 80%
of the intended value, however, release to the atmosphere was negligible
even when operating with only one bed in service. (BB-79)

The following is a list of items which were thought to have functioned
or are to be recommended.

(a)

Regular on-site planning meetings and communication in
general. (M-53)(T-66)

Cooperation between groups. (D-22)(I-36)

Daily reports. (K-U4T)

The numbering system. (K-43)

Change request system. (M-5L)

Combination heater and thermocouple drawings. (K-L42)
Microfilm drawing reader-printer. (I-37)

Operating procedures. (M-53)

Critical path scheduling of the fabricatiocns of components
and for doing meintenance. (M-54)(T-65)(T-66)

Tdentification and records of material used in construction.
(c-22)(R-63)

Preoperational and precritical testing of equipment. (N-55)
Operator training and reviews. (M-53)(0-56)

Use of a simulator for training. (K-45)(M-53)

All annunciators repeating in the control room. (K-kl)
Location of surveillance specimens. (C-22)(R-64)

Fuel and coolant circuleting pumps. (F-30)(Z-77)

Dust filters on ventilation system inlets. (L-4g)
Instrument maintenance. (D-23)

Seven days continuous operation of the computer before
acceptance. (Q-61)

Design data sheets. (K-L43)
well

(u)

14

The high quality construction used in the cover-gas and offgas
systems. (BB-T79)(BB-80)

The following is a list of items which were thought not to function
or were needed but not supplied.

Drawing index and cross-referencing. (A-16)(K-40)(A~17,19)

A well-organized written description of the control circuits.

(H-35)

Elementary drawings for the safety system, solid-state circuitry,
and the electrical system. (K-Lg)

Thermocouple wells in the fuel salt system. (E-26)(H-34)
Thermocouple wire penetration seals. (U-T3)

Two manipulators at the sampler. (V-75)

Drain tank weigh cells. (E-29)(H-35)

Reactor and drain tank cell reference volumes. (L-48)
Electrosystem temperature modules on freeze valves. (0-59)
Lube o0il system oil cooling. (F-30)

Facilities and storage for remote maintenance., (A-16)(A-17)

(A-17)(I-36)

Vent house remote maintenance. (A-1T7)

Building cranes. (A-18)

Remote maintenance jig and fixture program. (A-18)(Z-77)
A data-plotting board. (D-25)

A fuel salt flowmeter. (E-26)(U-72)

Radiator doors. (M-50)

Check valves. (0-59)

Punched card reader, line printer, and more magnetic tape units
for the computer. (Q-62)

Computer manufacturer's documentation. (Q—61)
Galloping safety circuits. (U-72)

Radiator stack flowmeter. (U-T2)

 
15

The following is a list of items which were installed and later found
not to be needed.

(a) Automatic load sequencing. (M-52)
(b) Thermocouple patch panel. (D-24) (K-43)
(¢) Helium treatment station. (L-48)(BB-T9)

The thermocouple scanner is the only item on which there was
disegreement. (D-24)(F-33)(U-71)

 
16

COMMENTS AS RECEIVED

In this section are reproduced verbatim the 28 replies to the

request for criticism of the MSRE. The identity of the writer is given
in Appendix II.

Remote Maint.

Remote Maint.

Organization

Remote Maint.

Documentation

Personnel

Reply A

Remcte Maintenance, or "We Sure Lucked Out on That One"

oomehow or other we were able to fix whatever broke
at the MSRE. Indeed there were times when further operations
depended upon successful completion of rather difficult
remote maintenance operations. However, these operations
were always slow and expensive. Therefore, the comments
that follow do not say that we did it wrong. To the contrary,
I think we did most of it right and well. The gist of what
fcllows says that we could have done it better and that these
points should be considered for the next time.

"Confusion Junction"

The "system used for remote maintenance" was a poorly
organized collection of parts, pecple, and procedures, that
was accorded a rather vague status in the organization of
reactor cperations, and was granted very little authority.
The responsibility, however, was clear and fixed. . '"Be ready
to fix it if it breaks."

Let me elaborate. The personnel involved with a Jjob
involved develcpment, operations, health physics, and craft
people. Information appeared on General Engineering drawings,
Reactor Division drawings, Luther Pugh's MSRE sketches,
Blumberg's and Shugart's sketches and drawings, two volumes
of photographs, and two loose-~leaf books of unpublished
procedures. The equipment to be maintained consisted of
all the equipment in four radicactive cells. The equipment
to do the maintenance was physically scattered all over "the
county." With each job the ground rules changed, although
this stabilized toward the later operations. During the con-
struction and precritical phases I had the definite feeling
that I was not needed and not wanted. This changed after
the reactor went critical. I was needed but still not wanted.
This was borne out by the marked difference in the effort required
to get craft work done between the time that I did remote
maintenance and the second gamma scan. For the latter Jjob
I was "on the team" and things got done very easily.

 
 

17
Reply & (Con't)

It is suggested that remote meintenance be made an Qrganization
integral part of the operations group. Remote maintenance
needs to be "on the team."

"Low Budget Productions, Inc."

Many of the decisions which affected my work were
based on "time and money shortages.'" It may be illuminating Remote maint.
if we enumerate some of the things we could have "bought"
had the schedule and budget allowed and then see what benefit
would have been derived. We could have: (1) trained Training
2 or 3 teams such as Shugart and Blumberg, (2) conducted
the P&E craft training program in the precritical period
when we needed it rather than later on, (3) provided better
physical facilities (these are enumerated below in terms
of defects).

The benefit from having provided the above items would
have been shortened downtime for the jobs that were done,
resulting in some cost saving, and a more positive demonstraticn
of the "maintenance capability.'" Further, if the project
had been requested to recover from the hole in the drain
system, and to put the reactor back on line, then an expenditure
such as the above would have been a good investment on
a cost basis.

"Nobody Knows the Facilities I've Seen"

The following is a list of the defects in the equipment
that we had to work with:

1. The vent house was never designed for maintenance Remote Maint.
of the radicactive components located there. There was
no head room, no overhead handling facilities, both the Des - Misc.

permanent and the maintenance shielding were crude, it

was crowded by the beams, pipes, electrical conduits in

the area (Some of these cut down the usable overhead access
area, the room was hard tc decontaminate, and the weather
inside the room was very closely coupled to the weather
outside the room.

2, There were never enough shielded carriers. We Remote Maint.
could have used more, bigger, and better "vitro" shields.
Usually we had to design and fabricate these during the Planning

heat of the shutdown.

3. The high-bay storage rack for remote maintenance
tools was toc small. Other equipment was piled on top Remcte Maint.
of the tools resulting in damage. Neighbors on both ends

 
Remote Maint.
Management

Des -~ Mech.

Management

Planning

Des - Mech.

Remote Maint.

Management
Remote Maint.

Remote Maint.

Documentation

 

18
Reply A {Con't)

of the rack moved in and restricted access to the ends

of the rack. A permanent, adequately sized area was promised,
but never delivered. Instead we received wit and humor

such as '"These tools make the high bay look like Big Foots
Used Auto Parts."

h. The portable maintenance shield was a very practical,
hard-working, essential piece of equipment. However, because
of budget restrictions, it was not of good gquality in many
respects. The main bearing was crude which made it hard
to turn. We could have saved set-up time with a system
of jack-screws for leveling the whole unit. Stronger, stiffer
tracks would have been helpful. The light and tool plugs
and penetrations were rough, and made some tool-handling
cperations difficult and time-consuming. The next time
around, this design should be used as a guide, but the
overall quality should be better.

5. Both building cranes were old, and frequently
required repairs, and usually at an inconvenient time.
Most of the trouble was electrical but we also had some
mechanical difficulties with the trailing cables. One
has only to examine the shutdown schedules prepared by
Webster and note how many of the critical jobs involved
either or both the cranes and the portable maintenance
shield, to realize how important reliability and good performance
is for this type of equipment.

6. The bolts used in the cell for remote operations
were poor in the following respects. The head size was
small in relation to the bolt diameter, thus diminishing
the bearing surface. The threads were rough and splintery
and as a result many bolts were discarded. Because the
bolts were expensive, not many were bought. For this reason,
we ran out and had to reorder a number of bolts and this
was expensive. There were at least two cases of galled
bolts in the MSRE cell that had to be removed and replaced
remotely.

T The jig and fixture program turned out to be
non-existent in-so-far as the items we needed to replace
such as the 522 and 523 spoolpieces and venthouse piping.

"The Information May Be on the Drawing —
If You Can Find the Drawing"

 

 

The job of remote maintenance involved a large amount
of reactor equipment that was described in the MSRE reactor
drawings. However, the information required for remote
maintenance was not taken into account enough in the preparation

 
19
Reply A (Con't)

of the drawings. The drawings were complete only in the
aspect that there was enough information to build the reactor.

When one considers maintainability, the following
information is needed.

1. Where is the item located with respect to the
cell block layout.

2. What is in the way or near by.

3. What clearances exist.

4,  What does it attach to and how.

5. What does the removable part consist of.
6., What are the envelope dimensions.

7. Finally (although not absolutely necessary) how
is it supposed to be replaced.

While all of the above would make it easier to do Documentation
maintenance or any other drawing research, I would like
to suggest the following minimum steps.

1. Reference drawing numbers and dimensions should
be used liberally on the face of the drawing.

2. A systematic method for going from any detail
tc the next larger assembly and so on and also the reverse
direction should be employed.

3 There should be more liberal use of drawing indexes
and/or picture drawings with the assemblies called out on
them.

"And As the Sun Sinks Behind the Decontamination Facility"

Let me now summarize my opinion on this subject. The
low budget policy was a realistic one, and turned out to Management
be correct. However, had the project been extended for
another fiscal year, it would have been a rather expensive
way to save money. Furthermore, for larger reactors with
their bigger costs and bigger potential hazards, we cannot
afford the luxury of saving time and money at the expense
of maintenance. We must overdesign, overbuild, and overprepare
instead.

 
Chemistry

Chemistry

Management

Chemistry
Planning

Sampler

20

Reply B

These comments are related to fission product studies,
and are generally related to two kinds of restrietion on
experimental flexibility: inadequate access to regions
and insufficient variation of conditions of chemical interest.
Some of the comments are a result of the slow development
of our understanding, some are related to design inflexibilities,
some to caution and the desire to demonstrate long term
smooth operation, and much to the priorities of other experiments.

As a matter of general philosophy, at some phase of
development of chemical machines, the effects of operation
under chemically extreme conditions should be explored
experimentally. This is preferably done in pilot plants,
such as MSRE or MSBE. This was done in MSRE mostly by
the inadvertent increase in oxidation potential of the
system by means presently subject largely to conjecture,
and the subsequent addition of reductant metals to the
salt to regain the desired though not determinable U®/U"
ratio. Argon replaced helium, purge flow rates were changed,
system flow rate was altered, and modest changes in operating
temperature and pressure were made, though these changes
were mostly organized for physics purposes. However, other
useful variations were not made. These include additions
of gases such as hydrogen, HF, oxygen, water, methane,
and possibly other gases which if added in judicious amounts
to the pump bowl purge could have had very interesting
and informative effects on oxidation potential, oxide behavior,
tritium behavior and other phenomena. To make this effective
would have required good analytical techniques for U3/U*
and tritium analysis (and sampling) among others; these
appear to have been nearly ready when MSRE operation was
stopped.

It was not possible to perform experiments in the
salt other than dipping a few materials in the pump bowl
liquid, and withdrawing for examination. In particular,
no electrodes, nc gas sparging, no filter collection devices
(except magnets) for solids or colloidal materials could
be effectively employed. This was in great part because
the only access was by means of the sampler-enricher system
to the pump bowl. The desired approaches would have been
feasible for an on-site hot cell containing a suitable
flowing salt bypass — admittedly difficult but essential
to the attainment of the data usually firmly stated to
be required.

The only access for fuel and gas samples was within
a spiral spray baffle, with ill-defined salt circulation,

 
21
Reply B (Con't)

undefined gas circulation, and essentially no possibility
of determining accumulation of materials at the gas liquid
interface. Contamination of sample capsule exteriors by
deposits in the transfer tube is strongly suspected but
poorly defined. Samples of pump bowl salt will represent
circulating selt reliably only for substances certain

to remain in salt during its sprayed passage to the sample
region. Next time, flow-through sample devices as & bypass
on a flowing line, if achievable, will provide samples

far more representative of the flowing salt.

The quality of the gas samples was difficult to assess
for several reasons. Samples were sucked into a capsule
in the space gsbove the pump bowl liquid within the spiral
spray baffle, with varying gas flow down the transfer
tube through this space. Transport conditions and residence
time for gas borne materials are difficult to define for
the pump bowl proper — though presumed "well-mixed" inspite
of baffles — but the gas situation within the spray shield
is even more elusive. Further, the effect of evacuation
of regions containing the seample capsule during removal
procedures is not clear.

The proposed off-gas sample station to have been
attached at the 522 Jjumper line could permit better samples,
but shut down schedules halted fabrication. Even better
would be & suitable flow-through rig capsble of regular
sampling.

Samples were not obtained from the drain tanks —
in particular, some ingrowing 25 Np appears to have been
lost there but periodic samples were not requested to
establish this. Flush salt samples similarly.

Removable metal specimens, in the heat exchanger
inlet and outlet regions for example, could have shed
much light on deposition of corrosion products and fission
products around the system, especially as g function of
altered temperature.

Reply C

One of the more serious deficiencies of the MSRE
was the lack of a routine carbon analysis on the fuel
salt samples taken to monitor corrosion products.
thecretically plus tests out of and in the presence of
irradiation indicated that there were no compatibility
problems under normal MSRE operating conditions. However,
with molten fluoride fuel salt circulating through the

Admittedly,

Chemistry

Planning

Sampler

Chemistry

Sampler

Chemistry

Planning

Chemistry
Planning

Chemistry

Chemistry
Qual Assurance

Des - Comp & Sys.

Organization

Operations

Management

 

22

Reply C (Con't)

approximately U4 tons of unclad core moderator graphite,

a major reactor component; it seems advisable to monitor

the carbon content in the fuel salt. The core surveillance
specimens samplings were spaced too far apart to adequately
monitor unusual operating conditions that might have developed.
I understand that an improved carbon analytical technigque
would have been necessary, but I believe that it should

have been developed and used as a routine corrosion products
monitor as suggested several times.

In the early part of the MSRE, fabricated Hastelloy N
components had the heat numbers (HT) and inspection record
(IR) numbers etched on every part. This was good and should
be practiced at all times for all MSRE materials. Toward
the latter part of the MSRE this practice seemed to decrease.

The core surveillance (= 3 in. from the vertical centerline
of the core) and the vessel (~ 5 in. from the exterior of
the core vessel wall) specimens were invaluable monitors for
the status of the Hastelloy N and the moderator graphite.
However, the cross sectional space limitations for the core
surveillance specimens complicated their design and hot cell
handling which ultimately restricted the desirable variety
of specimens.

Reply D

Comments on the Relationship Between Experimenter and Operator

Speaking as an operator I think the cooperation between
the experimenter and the operator was good. This is certainly
necessary if the maximum benefit is to be derived from an
experiment such as the MSRE. I want to comment on the occasions
which the experimenter and the operator should disagree for
the benefit of the experiment.

The experimenter should properly be aggressive and imaginative
in his requests for running tests of interest. The main
interest of the experimenter however lies in the results
of the test — not necessarily in the continued well-being
of the system.

The operator, on the other hand, should be encouraged
to adopt a possessive attitude toward '"his apparatus'" and
to resist any proposed test he feels may compromise its continued
well-being. This resistance should be encouraged and should
be viewed by management not as a lack of cooperation but
as a necessary function. This is not to suggest that operators
23

Reply D (Con't)

were ever accused of non-ccoperation. However, I do think
there were occasions 1t might appear as such unless this
point is considered.

For the operator to fulfill his role as the protector
of his apparatus means also that he oppose the violation
of a rule or limit for the purpose of obtaining data.

In addition to the possible mis-use of the equipment,

the attitude of the operator can be adversely affected

by unusual demands of the experimenter. (I speak now

of the "operator" as the man on shift who is responsible

for carrying out the instructions given him.) Perhaps

a large part of the problem is having the limits set unneces-
sarily close to begin with. I have had cperators comment

to me "we would have been fired for doing this (i.e.,
operating in this manner) last week."

The sense of participation and the attitude of the Organization
operator is very important to successful operations such
as the MSRE. During operator training, a large amount Training

of emphasis was placed on strict adherence to operating
procedures and limits. Yet (in the eyes of the shift operator)
the experimenter, apparently with very little difficulty,

obtains approval to seemingly viclate the most sacred

of limits. The operator thus very naturally loses some

respect for all limits. The fact that the change in limits
necessary for the experiment was given very careful considera-
tion before approval is beside the point. After so many

of these type changes it appears to the operator that

limits may be changed at will to accommcdate the desires

of the moment. One obvious solution of course is to have Management
the limits set at more reasonable values which will protect

the system and also accommodate mest of the foreseen experiments.

Instrumentation — General

As a whole the installed MSRE instrumentation ranged Des ~ Instr.
from adequate to excellent. The versatiliity and wealth
of information provided by the control room jumper board
is an outstanding example of instrumentation for the operator.
Certainly the maintenance people did an excellent Jjob Maintenance
of revising and maintaining instrumentation.

Along with some difficulties with installed instrumentation,
however, I want to mention the difficulties due to instrumentatiocn
not installed. There were many areas where, for purposes

of saving money, control and monitoring were left to the Des - Instr.
operator. OSome of these were important areas such as
the line 200-201 penetrations, the cell temperatures, Management

the FV-105-106 tee, freeze valve controls in general,
coolant salt flow transmitters, etc. Taken individually
Des - Instr..

Records

Des - Instr.

Des - Elec.

oL

Reply D (Con't)

or even in small number any of these could certainly be

taken care of by operator attention. However, taken collectively
there were sometimes too many situations and not enough operators
to do the job properly. More controllers would have saved

a lot of operator time and error.

Thermocouple Monitoring — Patch Panel

For critical areas such as the pump bowl, cell ambient
temperatures, line 200-201 penetrations, the FV-105-106
tee, etc., a better system of temperature readout and control
would have been useful. The thermocouple patch panel was
designed to provide a wide range of temperature readout
possibilities. In actual use, however, it turns out to
be a burden instead of a blessing to the operator. For a
simple thermocouple change requiring less than a minute
at the patch panel another ten minutes are necessary to
make 4 thermocouple log entries. The result was that sometimes
either needed changes were not made or a "temporary change"
was made in the heat of necessity and not logged. I think
that a patch panel connected in parallel to the normal thermo-
couple readout would be of more benefit. Logs would remain
necessary, of course, but only on the semi-permanent attachments.

Thermocouple Monitoring — Scanner

At the present time it appears the next reactor will
have a cell furnace type heating system and thus MSRE experience
would be of no benefit. However, perhaps there will be
auxiliary systems and lines which will be heated and monitored
individually. I recommend that a different type heater
and monitoring system be used. The oscilloscope scanner
system is satisfactory for meonitoring temperatures after
the system is hot but is almost useless for heatup. I propose
providing either more temperature recorders in close proximity
to controls or even better a scanning system which controls
individual heaters to a single, manually variable set point.
An entire system can be brought up to temperature uniformly
with one control. The autotransformers could be used for
trim to compensate for the difference between full rated
heater voltage and the voltage necessary to obtain the desired
temperature.

This same scanning controller system could also be
easlily adapted to maintain desired temperature gradients
within desired limits. This is an important feature when
heating long sections of piping which may contain frozen
salt or when minimizing thermal stresses due to steep gradients.

 
 

25

Reply D (Con't)

Control Room Layout

 

I suggest the design layout of the next contrcl room Des - Misc.
provide a wall space, convenient to both operator and
observer, for a data plotting board. There are numerous
times when a continuous plot of several variables is necessary.
This display of a few well chosen variables should be
made possible without cluttering up walls and windows.

Computer

The computer has been a valuable aid to the operation Computer
of the MSRE. From the operator viewpoint, the two most
exasperating problems were (1) "garbage" printout on the
alarm typewriter, and (2) an unnecessary multitude of
alarms due to an insignificant variable oscillating about
an alarm setpoint. I think the next use of a computer
for reactor operations should have its 1imit check duty
drastically reduced. It was intended that the computer
be an "early warning" device which would give alarm at
some value before the conventional alarm scunded. However,
operatcr confidence in the computer is lowered considerably
when trash comes out. Aggravation is compounded when
it rings a bell every time while doing it. When operator
confidence in the computer is lost, the value of the computer
to the project is diminished. A significant alarm signal
is easily lost when it is buried in a pile of insignificant
alarms. I suggest that the conventional alarm system
be allowed to monitor most of the variables and the computer
be used only for special items.

Sampler Enricher

 

In a manner of speaking the sampler-enricher was Sampler
more difficult to operate than the reactor. The reactor
itself was very forgiving of mistakes and would accept
an operator error with no adverse effects. The sampler, Organization
however, was very unforgiving. Many times it would respond
adversely even with the best of attention and care. The
sampler-enricher was, by necessity, a complex combination
of mechanisms. Operating check lists were eveolved which Operations
are probably the best that can be devised. However, it
is evident that detailed check lists can go only so far
with a complicated collection of mechanisms such as the
sampler. Continuous operator familiarity would be desirable.
I suggest that for the next reactor, manpower be distributed
such that a special crew performs all sampling operations.

 
Reactor Analysis
Management

Des - Instr.

Reactor Analysis

Des - Instr.

Reactor Analysis

Des - Instr.

 

26

Reply k

I think it is obvious that there are many features
and sub-systems associated with the MSRE that performed
their intended functions admirably. There seems to be
little advantage in dwelling on these since it would only
lead to a long narrative of doubtful value. However,
there are also some areas in which notable deficiencies
appeared. The following discussion deals with some of
the deficiencies that are apparent from my peint of view.

Instrumentation

 

I have now been closely associated with the operation
of two reactor experiments at ORNL; in both projects,
detailed analysis of the performance of the system was
significantly hampered by a lack of instrumentation. 1In
most cases this lack resulted, not from the fact that the
right kinds of instruments didn't exist, but from a failure
to design the right device into the right place. In fairness
to the designers, it must be admitted that some of the
needs could not be accurately predicted at the design stage.
Other deficiencies apparently resulted from conservatism
in the design to maximize system reliability. Maximum
reliebility may be an admirable goal in a plant that is
to do nothing but run for 30 years or so but, when this
goal compromises the ability to gain information from
a relatively short-lived experiment, it becomes a handicap.
Some of the specific deficiencies in the MSRE are discussed
below.

Primary Fluid Flow Rate — The lack of a flow element
in the primary loop has contributed much to analytical
difficulties. Such a device could have added a great
deal to our understanding of the system — it might even
have helped resolve our current uncertainty in power level.
The device in the coolant loop has demonstrated that such
an installation need not compromise the integrity of the
loop. A direct measure of primary flow should be strongly
considered in the next reactor design.

 

Temperature Indications — Accurate, fast-response
indications of the temperature of the primary fluid at
several points around the loop, and especially at the
core inlet and outlet, are essential to an accurage evalu-
ation of system performance. The thermocouple well in
the reactor neck (TE-RSE) was a move in the right direction,
but not far enough. Unfortunately, even this step was
nullified, apparently by an installation error. Since
there was only one thermocouple, the margin for error
was zero.

 
27

Reply E (Con't)

By far, the best temperature information obtained
ocn the MSRE came from the thermocouple wells in the coolant
loop. At this time there is no evidence to indicate that
any difficulty whatever existed with the wells. (Perhaps
these shculd be examined to confirm their integrity.) On
the basis of this experience, carefully designed thermccouple
wells should be seriously considered for several locations
in the primary loop of the next reactor.

Pressure Indications — This may be one area in which

 

the analytical needs (desires?) could not be accurately Reactor Analysis
predicted. However, we now know that a pressure-measuring
device, with good high-frequency response, at some point Des - Instr.

in the ligquid filled portion of the primary loop would
have been a valuable asset in the studies of bubble effects
and dynamic behavior. If such devices are not currently
avajlable for molten-salt applications, it would probably
be worth some development effort to get one (or more) for
the next reactor.

Computerization — It seems evident that the next reactor

 

will make extensive use of digital computers for collecting, Reactor Analysis
logging, and otherwise processing operating data; an on-
line machine might even be used for some contrcl functions. Computer

A number of the signals from the MSRE were not particularly

well adapted to computer handling — at least not by the

machine we had. I believe the entire area of compatibility

between the system instrumentation and the computer should Des - Instr.
be a prime consideration from the beginning of the instrument

design work.

In this same connection, noise analysis techniques
are now much more highly developed than they were when
the MSRE was designed and built. These techniques can
be used to quantitatively monitor the performance of rotating
machinery as well as for the more exotic applications
like neutron nocise — if the necessary hardware 1s provided.
We spent considerable time and effort setting up special
equipment in the blower house each time we wanted to check,
in detail, the bearing vibrations on the main blowers.
Consequently, these checks were made very infrequently
(and none at all were made on other devices). This information
could have been obtained routinely, and with better resolution,
by transmitting the vibratiocn signals to a computer with
adequate high-speed sampling capability. I believe the
next reactor should include a network that permits routine
noise analysis on every vital piece of rotating egquipment.
There may be analog devices that can do this for each separate
application but, if we have a digital computer with the

 
Operations

Planning

Management

Des - Comp & Sys.

Sampler

Des - Mech.

28

Reply E {Con't)

inherent capability and flexibility to handle this job along
with pressure noise, neutron noise, and all the other data
handling functions, we should take advantage of it. At the
computer end, the same basic hardware and software could

be applied to all noise signals which might make it relatively
cheap.

Calibrations — We found, when we finally needed it
and looked for the data, that some of our process radiation
instruments had not been calibrated in terms of absoclute
activity concentrations in the process streams. I realize
that such calibrations are difficult, and it may be even
more difficult to relate an activity reading on a sample
line to the real item of interest — activity in the process
system. However, some effort should be made to calibrate
each device on the reactor in terms that are physically
significant.

Control Rods

Although the control rods used on the MSRE worked out
all right, there is no doubt in my mind that life would
have been much simpler had we had ordinary, straight, rigid
control rods. I am not convinced that the problems we avoided
by restricting ourselves to one reactor-vessel penetration
outweighed the problems we encountered in squeezing everything
through that one penetration. While MSR's do not need flux
flattening for uniform fuel burnup or burnout protection,
there are advantages to be realized in terms of graphite
thermal stresses and lifetime. Since the next reactor will
need control rods, why not put them where they will do the
most good overall instead of basing everything on vessel
penetraetions?

Samplers

Considering the operating conditions and the number
of sampling operations, the MSRE samplers proved to be remarkably
effective and adaptable. In addition, the designers must
have learned a great deal about the various mechanical items
that will undoubtedly carry over into future designs. One
problem that may not have received adequate attention is
the possibility of cross contamination of samples in the
handling operations. It now appears that some highly sensitive
samples may have been affected by pickup of foreign material
in the fuel sampler. Future devices that are susceptible
to contamination by the materials that pass through them
would probably benefit from some provision for periodic
cleanup.

 
29

Reply E (Con't)

There were several occasions on which salt samples
from the drain tanks could have provided valuable information
about the state of the salt. A drain-tank sampler probably
would be a valuable research tool on the next reactor.
This is a good exsmple of an item needed for an experiment
that would probably not be desirable on the ultimate system.

Inventory Control

 

Accurate information on the salt inventory and its
distribution between the various parts of the system was
lacking throughout the operation of the MSRE. Some of
this problem can be attributed to the relatively poor
experience we had with the weigh cells. In retrospect,
bubbler tubes in the drain tanks might have helped considerably
since they worked so well in the pump bowls. However,
the piping layout, which allowed intermixing of fuel and
flush salts in poorly defined amounts also contributed
to the problem. The need for precise inventory data was
apparent in several areas of the analysis of system performance
and it will be required on any future system. The salt-
handling and salt-measuring systems should be designed
to provide the best information possible on how much of
what is where under all circumstances.

Unheated Lines

We learned the hard way on the MSRE that salt can,
in fact, get into gas lines appended to the liquid system
and freeze there. Thawing of these lines was a difficult
job on the MSRE because the piping layout made it hard
to heat the lines — even temporarily. We need to evaluate
every conceivable operation and misoperation on the system
and be prepared to handle the consequences of those that
are judged to be credible.

Record Samples

Experience with the MSRE has shown again the need
for keeping reference samples of all materials used in
construction of the reactor system. This principle was
carefully followed for the special materials (salts, graphite,
Hastelloy-N, etc.) but it was not followed as carefully
for the "ordinary'" materials. The question of Li concentration
in the insulation in the thermal shield illustrates the
need for keeping some of everything. Analytical results
are not necessarily required — these can be obtained when
and if the need arises, if the material is available.

Chemistry

Qual Assurance

Des - Instr.

Des - Piping

Des - Misc.

Management

Construction

 
Des - Comp & Sys.

Des - Comp & Sys.

Des - Comp & Sys.

Operations

Des - Elec.

Development

Des - Piping

Des - Comp & Sys.

30

Reply F

Lube 0il Systems

These "packages' with their "double redundancy" should
be very reliable to supply oil to future salt pumps of the
MSRE type. The Gulfspin-35 oil also gave remarkable service
in radiation fields.

The philosophy of having one pump of each package fed
from a semi-independent reliable power source is a good
cne.

Improvements in future heat exchangers for these type
of oil packages should be made. The next heat-exchanger
units should have a lot of excess capacity to allow for
loss in efficiency due to fouling from scale. Improvements
are also needed in cooling water chemistry, i.e., better
inhibitors tec prevent scaling.

0il should be completely changed after every two years
of operation as a matter of principle.

Lube oil pump motors for MSRE-type systems should
be the very best commercial grade of moisture and dust
proof design. These are inexpensive considering the job
they have to do.

Improvements in the priming of the idle pump of a package
are still needed.

Leak Detector System

 

It is recommended that this type of system be used
in all future plants which have "mske and bresk" fittings
and must be well contained. It is, however, economically
desirable, and now practically feasible, to greatly reduce
the number of these type of joints by the use of remote
welding and brazing techniques. This should be done wherever
possible for reasons of economy. It is obviously expensive
to provide buffered joints of the type used so extensively
at the MSRE.

Coolant Salt Pump

This coolant salt pump was an excellent piece of equipment
and performed far beyond its expected service life. The
fault of the oil leak to the pump bowl has already been
rectified for the future in that the spare rotary element
has a seal weld in lieu of a buffered metal "0" ring in

the critical area. This pump is recommended for future

 
 

31

Reply F (Con't)

Coolant Salt Pump (continued)

 

coolant salt circulation. Minor medification and shielding
may be necessary to pump sodium fluoroborate.

Freeze Valves

Try to eliminate via elevated or pressurized drain Des - Comp & Sys.
tank.

Put 2 or more FV's in series as a safety measure —
use alternately to reduce number of thermal cycles.

Freeze valves should have a tight fitting but unwelded
shroud. Could use braided SS cable for gasketing. A TC
on valve center could function with a heater on the air
stream to hold a setpoint with a minimum of thermal cycling
and could also help reduce thaw time by providing hot blast
air.

A freeze valve center might be a slightly reduced
section of INOR-8 pipe effected by spinning. This section
could then be annealed at ~ 1800°F before installation.

Future Fuel Salt Circulators

 

Mk II Fuel Pump — This is a fuel pump similar to Des - Comp & Sys.
that used at the MSRE, but with more height and volume
in the pump bowl (enough so that the 5.4-ft® overflow
tank could be eliminated). This pump is fully described
elsewhere and a prototype has been successfully run for
approximately 12,000 hrs in a test loop. This pump has
the advantage of a deeper pool of salt in which to spray
the fission product stripper flow, so that the tendency
tc such He gas into the pump suction is minimized.

 

This pump has the further improvement of a higher
space between the salt-gas interface and the point at which
the main offgas line leaves the pump bowl, hence the tendency
to plug the offgas line with spray mist and foam is decreased.

The Mk II pump is also seal-welded against oil leakage
into the pump bowl proper.

The fact that there is no overflow tank eliminates
the bothersome operational chore of burping the overflow
tank, which was at times of plugging in the offgas line,
a real hazard to the on-stream percentage time of the
MSRE,
Des - Comp & Sys.

Des - Comp & Sys.

Chemistry

Development

Des - Comp & Sys.

32

Reply F (Con't)

Future Fuel Salt Circulators (continued)

 

Any future reactor of this type should, as a bare minimum,
have this type of fuel pump.

Dual, Heatable Offgas Lines — Any type of salt circulator
built in the future should have at least two large, heatable
main offgas lines. These should alsc be fitted with very
coarse INOR-8 wool mist filters, and should be so valved
that one could pass the offgas flow while cold, as the
second was heated in quiescent state to melt the trapped
salt and allow it to flow by gravity back intc the pump
bowl.

 

Simplified Salt Circulator — A different approach
to the problem of fuel salt circulation than the traditional
"pump, pipe, and pot" would be to use a slow-speed pump
integral with the reactor. The use of bismuth based alloys
as a replacement for the coolant salt is also suggested.

Vacuum stripping of Fission Products — It has been
suggested that tests be run at 5 psia vs 20-psia pump-
bowl pressure to determine whether or not improved fission
product stripping could be achieved at the lower pressure.
Because of heavy existent schedules and an extreme shortage
of money for the total molten-salt program, this proposal
could not be tested in the MSRE. Because of its possible
beneficial effect on breeding ratios of future MSBR's and
because it would offer a system simplification if properly
designed, the idea should not be left untested if the molten-
salt program is carried on logically.

Partial Safety Drain

 

A partial drain of the reactor might be cbtained by
having the bottom of a tank (partial drain tank) connected
to the bottom of the reactor by large salt piping and having
the top of the partial drain tank connected to the top
of the reactor by gas piping containing equalizer valves.
During operation, the partial drain tank would be pressurized
with helium to force most of the salt into the reactor.
When a drain is needed, the equalizer valves would be opened
and some of the salt would flow by gravity to the partial
drain tank.

 
33

Reply F (Con't)

Simplified Heater Control

Use multipoint recorder-controller for controlling Des - Instr.
temperatures of fuel and coolant systems. Use reduced
voltage on heaters to attain long heater life (as in MSRE). Des - Elec,

One 24-point Brown instrument will scan and give 'on-

off" control to 24 heaters. These will "scan'" the 24
points in 1-1/2 minutes. The locad-carrying contactors
would be located outside the containment cell for mainte-
nance accessibility.

These instruments used to cost $3000 each in 1950,
and two can be economically used for each bank of 24 heater
zones; one being a spare tied in to a spare set of thermocouples.

This type of control will hcld the system at the
desired temperature automatically, under all conditions,
and even shut off power if one zone overheats.

Auxiliaries

Try to reduce auxiliaries so that personnel in power Management
utility field can take over this type of reactor some day
(i.e. use stainless steel lined bed rock for containment
and shielding. This might offer the possibility of simpler
expendable cores which would not be set up for remote
maintenance. In event of rupture in primary, or even
MCA, this lower priced, "mass produced" cell in its bed
rock "cave'" could be written off, the cell filled with
concrete and abandoned.

Use much better grade fundamental controlling instruments

and omit the computer "fine tuning". Des - Instr.
Reply G
We have investigated applications of noise analysis Reactor Analysis

at the MSRE since the reactor first went critical in 1965,

The MSRE allowed us to develop and demonstrate possible
applications of noise analysis for system parameter measurement
(void fraction, Be) and for malfunction detection (plugged
offgas line). However, the full capability of the technique
was not realized because the original instrumentation was

not specified with noise analysis in mind. Our experience

at the MSRE should help in specifying the proper sensors

for noise analysis in the MSBR.

 
Reactor Analysis

Des - Instr.

Reactor Analysis

Des - Instr.

Procurement

Des - Misec.

Des - Instr.

Management

Des - Mech.

Des -~ Instr.

3k

Reply G {Con't)

Specifically, it was found that the cover-gas pressure
sensor in the MSRE was too far removed from the pump bowl
(and core) to be reliable for making quantitative measurements
of void fraction in the core. The pressure sensors in the
MSBR should be as close to the core as possible and should
have a flat frequency response up to 100 cps.

We also believe that noise analysis of temperature
fluctuations may have yielded useful information if the
signal wire from the thermocouples in the MSRE had been
shielded. ©Shielded wire should be used for the thermo-
couples in the MSBR.

Reply H

Containment

Some aspects of the containment requirements and containment
testing requirements were not adequately considered in the
initial design of the MSRE. TFor example, the coolant cell
to high-bay pressure differential was never completely satisfactory.
Thermocouple headers and other penetrations were very difficult
to test and repair. The containment and containment-testing
criteria should be established early in the design phase.

Thermal Stress and Stress Concentrations

Much more careful consideration must be given to thermal
stresses, thermal fatigue, and stress concentrations in
future reactors. An indoctrination program for all project
personnel is needed to avoid these pitfalls. The leak in
FV-105 is possibly an example of this: The design contained
a root crevice subject to cyclic stress where the air shroud
was Jjoined to the flattened pipe section. Once made, this
design feature (error?) was overlooked by perhaps dozens
of people who would usually not have accepted such a design.

Thermocouple Wells

 

The fears in regard to thermocouple wells appear at
this time to be unfounded. Additional wells would have been
halpful at the more important locations in the fuel and
coclant systems such as the reactor vessel inlet and outlet
and at the heat exchanger fuel and coolant inlet and outlet
lines. Although the thermocouples in the MSRE have performed
exceptionally well, we still do not know the reactor inlet
temperature very accurately at other than zero power (see
Fig. 2, MSR-67-19).

 
35
Reply H (Con't)

Conventional Equipment

 

Some of the difficulties with the main blowers, component
cooling pumps, and perhaps other equipment illustrate the
need for a more careful evaluation of the design and application
of conventional eqguipment.

Control and Safety Circuits

 

Throughout the operation of the reactor, it was customary
to refer to one or more large stick files of circuit drawings
or diagrams for any questions in regard to the various control
or safety circuits. A well organized description of the
various control and safety circuits and actions would
have been valuable for the initial operator training and
for continuing reference (i.e., what circumstances will
cause a rod screm, a load scram, a drain, . . . , high pump
pressure causes . . . , etc.).

Operational Errors

 

Of the few events in the reactor operation that can
be regarded as operational errors, the most serious was
probably the overfilling accident in July 1966. Although
the experiment that was being performed could have been
completed successfully with the quantity of salt that was
avalilable using different procedures, the accident would
have been avoided by increasing the loading of flush salt.
The loading equipment had been removed by the time the
problem was recognized, and a considerable amount of time
and effort would have been required to add additional salt.
It is obvious now that this time and effort would have
been well spent.

Drain Tank Weigh Cells

None of the drain tank weigh cells have given the
desired accuracy in regard to absolute calibration or to
repeatability. This could stand some investigation to
determine if there is something wrong with our installation,
if our weigh cells are below specification, or, if our
desired accuracy was Jjust wishful thinking.

Des - Misc.

Documentation

Training

Operations

Management

Des - Instr.

 
Sampler

Management

Sampler

Planning

Des - Mech.

Documentation

Management

36

Reply I

Sampler-Enricher

 

The space available for the installation cof the sampler-
enricher was limited, as so often happens in using existing
facilities for new equipment. Therefore, the size of all
containment areas had to be held to a minimum which crowded
the piping and left very little storage room inside the
sampler for tools and extra capsules. The crowded conditions
also increased maintenance problems.

The sampler-enricher was designed to take a 10-gram
salt sample and to add about 110-grams of enriching salt.
It was also used to (a) take 50-gram salt samples, (b) trap
20~-cc off-gas samples, (c) expose graphite to the salt
and off-gas for short periods, (d) study fission product
deposition on various metallic surfaces, (e) adjust the
Ut3/u** ratio, and (f) perform several other activities.
The design considered the enriching capsule as the maximum
size object to be handled, thus limiting the dimensions
of other devices that were used. For equipment that pro-
vides access to important areas of an experimental reactor,
such as the fuel and cover gas, functions other than the
primary one should be considered during the design. On
several occasions openings had to be cut through containment
vessel walls for modifications and repairs. Providing
extra access areas initially could reduce the necessity
for field alterations.

As noted above, many different types of capsules were
used during the time the reactor operated. They were developed
through the cooperation of the experimenters and the operations
supervision. This joint effort allowed many experiments
to be performed successfully without endangering the future
usefulness of the sampler.

Where shielding is required, the use of round or rectangular
shapes for equipment support members in place of angle
iron would reduce labor charges for installing the lead
without appreciable change in shielding quality or material
cost.

The photographs made of various components were helpful
in planning maintenance work. However, additional pictures
made from other angles would have helped in tracing lines
and determining the location of equipment.

For any item that must be decontaminated before it
can be reused, the cost of decontamination versus the cost

of a new part should be analyzed. 1In many cases it is

 
37

Reply I (Con't)

cheaper and better to use new parts rather than decontaminate
and repair used ones, as was found for the sample transport
container.

All piping that connects to the gas space of the Des - Comp & Sys.
pump bowl should be equipped with heaters for a distance
of several feet from the pump for removing unexpected

salt plugs.

Every member of the operations crew was trained to Sempler
operate the sampler-enricher so that a sample could be
obtained at any time without dependence on any one individual. Operations
However, several weeks often elapsed between times that
a given person would take a sample. Operations requiring Organization

the use of the manipulator and other special items were

more difficult when done only occasionally. This resulted

in more cperating errors and increased the required maintenance.
When everyone uses a piece of equipment, no one feels
responsible for it. When only a few individuals are assigned
to it, there is more personal pride involved for that

is "their job". Part of the procedure — isclating the

sample — could be done by anyone who followed the detailed
procedures. Remcving the sample regquired more operator
skill. For this operation, the use of only a few people
would have saved time and reduced wear on the equipment.

General

The microfilm reader-printer and the file of microfilm Documentation
drawings of all reactor components was useful and time-~
saving on many Jjobs. Also, prints were always available.
These files were maintained with the latest drawing revisions
by the reactor personnel.

The planning meetings helped to keep abreast of current Communication
activities at the reactor and of the immediate future
plans.
Reply J

Thank you for the opportunity to discuss what may
be learned from the design, constructicn, operation, and
maintenance of the MSRE. Although the remarks that follow
are concerned with deficiencies, please accept "Well Done"
to you and your staff for the very satisfactory and successful
operation of the MSRE. Among many other things, you have
handled the care and operation of the salt pumps in excellent
Procurement

Procurement

Management

38

Reply J (Con't)

fashion. It is difficult to conceive that even with my
personal involvement with the pumps, I could have done
sc well for them.

My remarks concern (1) the specification of, and the
selection of a manufacturer, for the drive motors for the
MSRE fuel and coolant salt pumps, (2) the leakage of oil
from the catch basin into the fuel salt in the fuel salt
pump, and (3) the plugging of the off-gas line of the fuel
salt pump. However, only the first item will be covered
here. The comments in the letter of January 28, from P.
G. Smith to R. H. Guymon, cover adequately the other two
items.

The contract for the four (L) electric drive motors
for the MSRE fuel and coolant salt pumps was signed in May
1961, and the fourth motor was delivered sometime in the
spring of 1965. This period of nearly four years is somewhat
unusual for the manufacture of T75-hp electric motors. The
fabricator of the hermetic vessels in which the motors
were installed must have taken a considerable '"bath" during
the process. Although it is not practical to summon up
all the reasons connected with the performance, two of
what seems to me to be the principal reasons do come in
mind.

1. During the survey for a suitable motor manufacturer
the emphasis was placed on the electrical engineering aspects
of the job. The selected manufacturer, Westinghouse of
Buffalo, New York, is a very competent designer of electric
motors, but had little or no experience with the design
and fabrication of ASME vessels and apparently none with
nuclear vessels. In its search for some expertise in this
field, Westinghouse selected Emil von Dungin Company, Lockport,
New York, to design and fabricate the vessel. It turned
out that von Dungin was in reality a manufacturer of decorative
iron work and of shipping containers for automotive radiators
and had very little experience with ASME vessels. With
the passage of time it became necessary to maintain an ORNL
resident inspector in the Buffalo and Lockport area to
monitor and assist the progress of the fabrication of the
vessels. I believe we maintained him for two or more years.

Hindsight indicates that the electrical engineering
aspects of the job were quite normal and that emphasis should
have been placed on the requirements for the fabrication
of the nuclear reactor grade hermetic vessel, Hopefully
in the future we will provide the survey emphasis that
is appropriate and adequate for the Job.

 
39

Reply J (Con't)

2. During a revision, perhaps the final one, of
the specification for the drive motors the vessel material
requirement was changed from stainless steel to carbon
steel probably on the basis of the economy to be expected
with the latter material., Several serious, and possibly
gratuitous, fabrication problems for von Dungin which
were caused by the decision include: (a) The difficulties
that were associated with the purchase of new material
from suppliers to the requirements of the low transition
temperature material specified by the Code and the later

attempts to upgrade material on hand. (b) The time-consuming

and expensive attempts to weld repair the "tears" in the
cylindrical vessels caused by the "weld buttering" that
was required to provide for a 'flat head" vessel closure.
(c) The many and expensive attempts to braze stainless
steel tubing, which was required for water cocling the

motor, to the carbon steel vessel. Although the coefficients

of thermal expansion for the two materials are not suited
to brazing, attempts were made to furnace braze and to
hand braze the coil to the vessel before they were finally
Joined with an electric arc process.

Hopefully, in the future we will exemine our decisions

for economy sufficiently to uncover some of the anomalies
that may prove to be expensive both in time and money

to us and hurtful to small businesses or small manufacturers.

Reply K

(1) In any project it is desirable to have all needed

information readily available. As the project gets larger
and more complicated, the need for formal documentation
becomes apparent. The more documentation accumulated,

the more difficult it is to keep all of the information
accurate and up to date. So at the MSRE we had reports
which described in authoritative language what had been
planned or perhaps what had actually been installed but

not what was currently being used. This caused considerable

confusion especially to newcomers but often also to old
timers.

It would have been nearly impossible tc have kept
all MSRE documents up to date. However, master copies
of certain key documents should have been continuously
updated. Periodic republication would have been desirable.

Procurement
Materials

Management

Management

Documentation

Management
Documentation

Management

Documentation

Management

Documentation

Management

Documentation

Documentation

Lo

Reply K (Con't)

(2) A design report should be written as scon as pcessible
(this includes all sections) and, as mentioned above, should
be kept up to date throughout the construction and operaticn.
The criteria used, reason for the design, and detailed
calculations should be kept in a central file. One possible
way of implementing this is as follows:

Appoint an editor (or editors) to write the final report.
He would set up a file for each system, component, and
instrument. He would approve all construction drawings
and purchase orders to assure that all information has been
submitted to permit writing the final report. Each designer
would be required to submit his calculations, criteria,
and reasons for his design as well as drawings before approval
for construction. Manufacturer's literature and operating
instructions would be required before a purchase order
would be issued. The design files would be kept at the
design office until the start of construction when they
would be moved to the construction site. Later they would
be transferred to the ocperator's office.

As an example of where this might help: the design
of the instrument alir system was not covered in any section
of the MSRE design report. This is probably the result
of 1t being designed by another design group or perhaps
it was not decided whether it should be included in Part
1 or II. The above system should eliminate omissicns such
as this. ZEach designer should be required to assure that
his system cor component is adequately documented.

(3) Drawings which need to be kept up to date should
be so designated and someone should be assigned to do this
(perhaps the design report editor). All project personnel
should be held responsible for advising him of any changes.

(4) All field changes should be transmitted in writing
t0 those responsible for updating documents. Most field
changes should be approved by the designer. The reasons
for each change should be explained.

(5) A drawing index with considerable cross-referencing
would have been helpful during construction and subsequent
operatiocn.

(6) The electrical elementary type drawings could
stand considerable improvement. These should be simple
enough to allow quick reference during operation but should
show all important information. Some things which should
be included are: physical location of switches, instruments,

 
L1

Reply K (Con't)

and equipment, all ammeters, voltmeters, and other instruments,
control circuit numbers, and control power necessary to

actuate switches. All switches, instruments, transformers,

and other equipment should be numbered. Control circuits

and switch tabulations should be furnished.

(7T) Elementary type control circuit drawings should Documentation
be developed for showing systems using solid state devices.

(8) The safety instrumentation elementary drawings Documentation
were inadequate. Engineering elementary type drawings have
not yet been issued showing that power, period, and temperature
will cause a rod scram and indicating the matrix system
used. Block diagrams were issued showing this about 2 months
before initial criticality. Period safety scram is still
nct shown on these. Documentation of the nuclear and safety
instrumentation for use by the operator needs to be given
high priority.

(9) Proper construction is, of course, a prerequisite Documentation
to successful operation. Therefore, drawings and other
documentation should be provided to assure adequate construction,
however, other documentation is needed for operation. This
does not need to be nearly as detailed but needs tc be
in a form which can be referred to repeatedly without undue
loss cof time. Some specific examples are given below.

(a) Flowsheets should be similar to those issued
for the MSRE. They should include line numbers and names,
equipment numbers and names, simplified instrument numbers,
locations of lines and equipment, direction of flow, normal
flow rates, normal temperatures, normal pressures, relative
elevations when important, and normal and fail position
of valves.

These should be reduced to 11 x 17-in. size. To avoid
unnecessary confusion, most data on components, cross-
references, and apprcval signatures could be on a separate
sheet or on the back of these. Such things as line sizes,
the number of thermocouples or heaters on a line or equipment,
flanges and leak detectors, would not need to be included.

The MSRE flowsheets were made so that lines extending
beyond the edge of one drawing matched the next drawing.
In other words, it is possible to fasten the flowsheets
together and end up with a large drawing of the entire
plant. This is desirable as it facilitates tracing lines.

 
Documentation

Documentation

Management

Lo

Reply X (Con't)

(b) Although flowsheets, as described above,
supply most of the day-to-day needs of the cperators, there
is often a need to know more detailed flowsheet type information.
Drawings showing these should start with the simplified
flowsheets so that the orientation on the drawings is the
same. Details of instrumentatiocon loops, like given on the
instrument application diagrams, should be added along
with any instrument valving and valve numbers. Control
and annunciator circuit numbers should be given from all
sensing elements as well as to all valves, motors, and
equipment. Line sizes and flanges should be shown along
with the flange leak detector lines. Approximate location
of thermocouples should be indicated. These drawings would
be the starting point for most trouble-shooting.

(c) Four types of control circuit drawings were
provided for the MSRE (block diagrams, engineering elementary,
maintenance elementary, and jumper board drawings). No
one type 1s entirely satisfactory for operations. The
block diagrams are probably the easiest to use when learning
the circuits but they are not sufficient for operating
the reactor since they do not readily show all the control
actions that occur, the switches that initiate the action,
relay numbers, jumper board lights, ete. The engineering
elementary drawings supply most of the information necessary
but are difficult to use. The jumper board drawings lack
detail and cross-referencing. With a little more effort
these three could be combined to give a set of drawings
which should be much more useful to the reactor operators.
These would be similar to Fig. 1. Switch setpoints should
not be shown but should be tabulated similar tc the MSRE
Switch Tabulation with provisions for keeping a master
copy up to date,

(10) Combination heater and thermocouple drawings as
used at the MSRE were very helpful. These should show
the approximate lccation of the heaters and the thermocouples.
Isometric drawings could be used to advantage in some places.

(11) Different groups often work on the same type
design or development. The documentation should be coordinated
so that the same set of drawings, the same reports, and
the same type nomenclature and documentation are used.
The same units should be used throughout the project (°F
vs °C). For instance, the control and annuncistor circuits
for the diesel generators are not shown on any elementary
type MSRE drawings. (Wiring drawings must be referred
to.) The circuits for the 13.8-kV auto-transfer switches
were only shown in the menufacturer's instruction drawings

 
L2a

18

]: 534 Emergency Drain Switch
S3C (on console)

 

I1SS-600B
N Open when FP QOverflow
}C)\ Tank level is high

K224 Open when FP pressure is high

‘ / IK261+A I KobA  Open when cell air activity is high
SAFETY"“YEE_:I:_______ég:
JUMPERS___Z K26A

 

 

 

—y = k2654 Open when coolant offgas activity is high
- T
ST
N/ K1ICT— K2C
< : Kl K3C ¢ Open when reactor outlet
. oD E3D temperature is high
N~ T I
/N ‘
)
S ka8 KB18

 

 

20,18 677,18

119,A655 11
108,118 T

 

EMERGENCY FUEL DRAIN
WHEN DEENERGIZED

 

ECC DESCRIPTION
20 Fuel Fill Restriction
117 Opens FD-1 Vent
118 Opens FD-2 Vent
119 Opens FFT Vent
A655 Thaws FV-103
677 Thaws FV-105
1084 Annunciation (4012-1)
Emergency Fuel Salt Drain
(MB-8-1)

 

 

 

Fig., 1. Suggested Type Drawings for Control Circuits

 
L3

Reply K (Con't)

until August 1967 when these were transferred to an ORNL
drawing. Neither are described in detail in the design
reports. The circuits should have been redrawn using

the same criteria as used on other engineering or mainte-
nance elementary drawings.

(12) Design data sheets as published for the MSRE Documentation
were very helpful. These should be expanded and kept
up to date. Management
(13) Design and operating criteria should be established Management

early. All staff personnel in design, procurement, construction,

analysis, maintenance, and operations should be thoroughly

familiar with these and should assure that their employees

know the reason for each piece of equipment or system

for which they are responsible. At the MSRE, the containment Containment
philosophy was not adequately specified. There are places

where all welded stainless steel systems tie into screwed

black iron systems.

(14) More planning should be done so that periodic Planning
tests, such as those made on containment or instrumentation,
would not be too time-consuming. On a power reactor, Containment

shutdown time is expensive.

(15) During the design of the MSRE, it was thought Des - Instr.
that extreme flexibility would be needed for interconnection
between all thermocouples and their readout instruments.
During operation, it has become obvious that the flexibility
provided by the patch panel was not necessary for a good
part of the thermocouples.

(16) The system used in selecting instrument, heater Documentation
line and valve numbers at the MSRE was an aid to training
and operation. ZEach system was assigned a series of numbers
(i.e., fuel system is 100 to 199). Any valves in a line,
instruments attached to the line or heaters installed
on it were numbered accordingly. (I.e. HV-522A is a handvalve
in line 522, PS-575C is a pressure switch attached to
line 575, H-101-2 is a heater on line 101, and TE-103-
1 is a thermocouple on line 103.) Due to modifications,
we ran out of numbers in some systems. It might have been Management
better to have assigned 1000 - 1999 to the fuel system,
5000 - 5999 to the cover-gas system, etc.

(17) Elementary drawings such as flowsheets, circuit
elementary and electric one-line drawings should always
precede detailed drawings such as piping and wiring drawings.
The elementary drawings should be approved by operations Management
personnel. DNumbers used on the elementary drawings should
be shown on the detailed drawings.

 
Construction
Operations

Des - Instr.

Des - Elec.

Des - Instr.

Management

Management

Operations

Organization

Operations

Ly

Reply K (Con't)

(18) Operations personnel should follow construction
to assure that it will operate. If not, construction personnel
should be thoroughly familiar with operating criteria.

(19) Pressure instrumentation on the primary system
of the MSRE was designed for operation and performed well
for its intended purpose. However, when the system was
opened for maintenance, it was desirable to exactly match
the cell and system pressures to avoid outleakage of activity
or inleakage of moisture. Instrumentation should be provided
to adequately monitor shutdown as well as operational periods.

(20) As at the MSRE, all annunciators should repeat
in the main control room. An improvement over our system
would be to have the control-room annunciator light remain
on until the remote difficulty is corrected. It should
not interfere with other alarms which are common to this
particular control-room annunciator.

(21) Safety actions were often in a two-out-of-three
circuitry. To provide more reliability, the three channels
were usuelly distributed between 48-V DC, 250-V DC, and
TVA (diesel) power, However, these were then fed through
a common circuit which nullifies the advantage. For instance,
low cover-gas supply pressure opened contacts in ECC-L46,
L7, and 48 which were on L48-V DC, 250-V DC and TVA power
respectively. Therefore, failure of one power supply would
only inactivate one third of the relays. However, these
were fed into matrixes in ECC-40 and 41 which were powered
by 48-V DC. So failure of 48-V DC power closed the block
valves. Therefore, ECC-46, 47, and L8 might as well have
been supplied by L48-V DC power.

(22) Detailed procedures for instrument checkout and
startup should be written during design as well as procedures
for replacement or repairs and subsequent checkout. These
should be written and updated by the instrument engineer
who will be assigned to the reactor operations.

(23) Operating procedures should be written in as much
detail as possible as design progresses to assure operability.
This 1s especially important on items which must be tested
periodically such as containment check valves.

(24) All design should be such that the equipment can
be operated by rotating shift personnel. If this is not
possible and individual attention is required, the project
should provide special operating personnel.

 
L5

Reply X (Con't)

(25) Plans for shutdown and decommissioning should Planning
be considered in the original design.

(26) At the MSRE, there were various times when location Des — Misc.
of instruments and equipment were shown on plan drawings
of the building (intercom station, beryllium sampling, HP
instruments, etc.). Each time this was done, a new drawing
was made of the building. It seems that this and other
similar duplication of effort could be eliminated.

(27) During preparations for operations, a table was
made listing all known variables in alpha numerical order. Operations
Blanks were provided for indicating the readout instrument
number and location, the function of the variable, the
normal value, the alarm and control setpoints and prcoposed
logging frequency. Members of the design and development Planning
groups were contacted to determine recommended initial conditions.
This provided a very good starting point for making up the
various logs and aided in writing operating procedures.
Design and development personnel should be responsible
for advising operations of any recommended additions or
changes in these.

(28) The operating crew should be at the site long Operations
before the cells are closed so that they can become familiar
with the physical layout. Training

(29) The use of simulators for operating training
was found to be very useful. The value of this was enhanced Training
by having the cperator sit at the conscle, use the actual
control switches, and refer to the regular reactor instrumentation.
Perhaps additional training could be done by simulating
failure of some of the control and/or safety instrumentation.
By running tests where the system is maloperated, the simulator
would very dramatically illustrate the consequences.

(30) An operator training manual should be written.
This would have been helpful in the early training and Training
would have saved considerable time in training replacements.
This manual should be updated as changes are made. The
reactor design, instruments installed, and methods of operation
should be covered from the operator's viewpoint. To aid
in learning and in handling emergencies and unusual occurrences,
the reason for the varicus features should be stressed.
Instructions should be included for manipulative training
and for aiding the trainee in becoming familiar with the
physical layout. Photographs of operating areas would
be helpful. Detailed outlines of tours and oral instruction
should be provided. Operating techniques such as changing
chart paper and log-taking should be included. The manual

 
Training

Computer

Documentation

Management

Records

L6

Reply K (Con't)

should be designed for the least adept operator and should
advance this person to the stage needed to operate the reactor.
Additional sections should be provided for the supervisors
including the art of supervisicn.

(31) The instructors used for training the MSRE operating
crew were engineers and scientists who, in most cases,
were real familiar with their subject. However, some did
not know what was needed by the operators and most had no
experience as teachers. It would be better if fewer instructors
were used and have the experts on hand to answer gquestions.
These instructors should be given some prior training on
such things as theory of learning, lesson planning, and
methods of presentation.

(32) The capabilities of the computer were not fully
utilized by the operating crew. I am sure that the attitude
of shift personnel and operating supervision played a part
in this. Insufficient knowledge of the computer may have
contributed. This could have been due to inadequate training
or perhaps inadequate emphasis on the need to use it. Unreal
or unnecessary alarms caused annoyances and lack of confidence.
The fact that conventional instruments were used for the
actual operation of the reactor and that most of the important
cperating signals had conventional readout instruments
reduced the urgency for use of the computer. The personnel
who were responsible for procurement and initial programming
were either not familiar with its potentials or with the
needs of the operators. With the experience gained at
the MSRE, we should be able to improve on design and utilization.

(33) Operating procedures should really be facility
procedures with sections for operations, maintenance, etc.
Personnel in each group should be responsible for learning
their section and keeping a master copy up to date. The
importance of keeping the same operating crew has been
proven. This applies alsoc to maintenance engineers, foremen,
and craftsmen.

(34) Considerable time was lost in trying to determine
what an oddly-positioned pipe was in the reactor cell.
It turned out to be a remote maintenance tool. There is
no record of when the tool was dropped. It is hard to admit
one's mistakes but it should be emphasized that this must
be done so that cothers can profit by them and so that necessary
corrective action can be taken.

 
b

Reply K (Con't)

(35) The daily reports were very helpful. I believe
that these should be expanded to include phases other
than operations. Often operational records were insufficient
but when the daily report was written, it was possible
to get the needed information. Often maintenance records
were inadequate and now it is impossible to determine
Just what was done. If maintenance were included in the
daily reports, this information could have been obtained.
Sunmary of analysis of data should also have been included.
Too much information may at times be cumbersome but information
not recorded may be lost forever.

(36) The salt and fission product chemistry was closely
followed at the MSRE., However, more information should
have been provided as to the lube-o0il chemistry, water
treatment, etc.

(37) Programmed maintenance of equipment as used
at the MSRE is desirable. However, the schedule should
be reviewed regularly by someone who knows the maintenance
requirements as well as the usage of the equipment. Results
cf the maintenance should be reported and reviewed.

(38) Lines, valves, wires, equipment, etc., should
be labeled as they are installed.

(39) The model of the reactor cell was very useful.
A model should be made of all areas which are not accessible.
All equipment lines, etc., should be labeled.

(40) Graphi~ panels showing important flowsheets
with valve positioun and equipment status and jumper boards
showing important control circuits and their status are
very helpful to operation.

(41) Repeating one of the critiques from the HRT,
do not use an existing building. If it is necessary to
do sc, drawings should be made showing as-built ccondition.
For example, the electrical system at the MSRE was left
over from the ART. Drawings were issued to show the changes
to be made. These were satisfactory and probably necessary
for construction, however, it was difficult to understand
the overall setup when one drawing showed what used to
be and another showed what was changed but no as-built
drawing was available,

(42) Another HRT item which is still applicable is
to have in-line or at least on-site analysis of samples.

Records

Chemistry

Maintenance

Documentation

Management

Des - Instr.

Management

Documentation

Sampler

Chemistry

 
Containment

Des - Instr.

Des = Comp & Sys.

Des - Piping
Procurement
Des -~ Instr.
Procurement

Des - Instr.

Procurement

Des - Piping

L3

Reply K (Con't)

(4L3) Two potential hazards existed with the containment.
I do not have any suggestions as to how to circumvent these.

(a) Often it was desirable to open the cell as
soon as possible after shutdown. At this time, the amount
of fission products were less than during operation and
the chances of a gross leak from a drain tank was probably
less than from the circulating loop. However, if conditions
similar to the MCA did occur, activity could have been
forced through the cell opening. The cell ventilation valves
were normally kept open to prevent this.

(b) It was not possible to weld the membrane
in place with the ventilation valves open. Therefore,
at the end of each shutdown, there were periods when the
membrane was partly or completely welded in place and some,
if not all, of the blocks were off. If at this time conditions
similar to the MCA had occurred, the cell pressure would
have increased and the membrane could have burst.

Reply L

Cover Gas

1. O2 and moisture analyzers should be located in
an area where the temperature is held constant.

2. Due to the purity of bottled helium, the treating
station is not required.

Containment

1. Pipe connections should be provided to test block
valves without disconnecting fittings.

2. Better distribution of temperature compensating
reference volume is needed.

3. Block valves to cell pressure and reference volume
should be on separate header from other block valves so

that the hook gage can be used during pressure tests.

Component Cooling System

 

1. Larger pipe diameters are needed for less pressure
drop.

 
L9

Reply L (Con't)

2, Better blower pressure relief valve or a combination Des - Piping
pressure relief and unlcading valve is needed so that blowers
could be started with no back-pressure.

3. Blower discharge check valves should have metal Des - Piping
"hinge" so flexing would not cause failure. Check valve
design should be of type or sized to prevent chatter and
wear.

L, Flow meters are needed for trouble-shooting Des - Instr.
(loss of capacity).

Water System

1. Insure that the design pressure of all components Des - Comp & Sys.
exceed shut-off pressure of the pumps.

2. Design the necessary parts of the system as secondary

containment and thus eliminate block wvalves. Procurement
3. Provide liguid chromate for cooling tower system

to prevent sludge buildup. Use a pump to meter inhibitor Des - Comp & Sys.

into system.
Ventilation System

1. Provide duct filters on inlet ducts to protect Des - Comp & Sys.
stack roughing filters.

 

Reply M
General Remarks
When I locok back on the MSRE, the dominant impression General
is one of excellence and success: in its design, development,

construction, operation, and maintenance. Of course as

I approach the task of writing a critique, a few shortcomings
and failures come readily to mind, and a quick scan of

the MSRE history brings back memories of a host of relatively
mincr difficulties. 1 believe that the performance of

the equipment and the mechanical problems that were encountered
have been adequately chronicled in the MSRP progress reports
and the summary reports which are now being completed.
Certainly the intention and the effort in this reporting

was to be quite candid, and I don't believe there is any
hitherto secret problem to be disclosed in this critique.

Nor is it practical to make the critique a complete recapitu-
lation of previously reported problems. One purpose of

 
General

Design

Development

Development

Operations

50

Reply M (Con't)

the critique 1s perhaps to hand out well-earned praise

where it is due, but again this cannot be comprehensive,

and those who contributed to this remarkably successful
reactor experiment have, I am sure, derived their satis-
faction and sense of accomplishment from actual events rather
than what may be written here. What then is the value of

the critique? Hopefully it will bring back to us and transmit
to those who follow us some lessons worth remembering, some
pitfalls worth identifying. I would hope that if you are
responsible for any aspect of the molten salt reactor effort
to which the MSRE experience is pertinent that you will

not only study this critique but will take the time to reread
the relevent portions of the progress reports and the summary
reports on the MSRE and make the next reactor even better.

Design and Development

 

The startup of the MSRE attested to the care in its
design and development, for there were few problems. One
startup problem revealed a design deficiency that was solved
by another trip to the drawing board: that was the bowing
end warping of the radiator doors. The original doors quickly
became absolutely inoperable, but adequate consideration
of thermal effects on the structure led to a quite satisfactory
redesign.

Improvements had to be made in the radiator door camming
and sealing and the leakage of the high-~-temperature air
through the enclosure box itself, reflecting what was in
effect a decision to complete the development of this component
after its installation in the reactor. This was costly
in terms of program delay.

The plugging in the offgas system presents at least
a couple of lessons. One is that behavior which may appear
first as only a small, sideline effect in a development
test aimed primarily at measuring or proving other things
should be very carefully evaluated, for it may become a major
problem in reactor operation. Keeping a system going by
rapping on clogged filters or valves or torching a line
plugged with salt spray isn't too much trouble under non-~
radioactive conditions, but may be impossible later. We
did this in the prenuclear testing of the MSRE and were
still doing it years later on the Mark-II fuel pump. One
should not rely on makeshift procedures, but should develop
a solution that will work in the reactor. The other lesson
is to anticipate that there may be more crud in a piping
system than you can logically predict. Flow restrictors
and valves, if they must be sized small for proper control
mist be protected by amply large and effective filters.

 
51
Reply M (Con't)

Extreme or unique cases of crud, such as the polymerized Des - Comp & Sys.
0il in the MSRE offgas, can't always be anticipated, but
there ought to be some allowance.

The importance of strong, closely coordinated development Development
support was evident in the response to the offgas plugging
problem, which produced in a short time the effective particle
trap.

The well-publicized limitation on the heat removal Chemistry
capability stresses the importance of having reliable
data on the properties of the salts, of special care and
adequate contingency factors in unusual designs, and the Des - Comp & Sys.
need for complete review when design criteria change.
(I understand that when the radiator was designed, the
MSRE power was expected to be 5 MW. The radiator was
designed for a nominal power of 10 MW, which was an ade-
quate way of insuring a 5-MW capability. But when the Management
reactor design power was raised to 10 MW, the nominal
10-MW radiator wasn't enlarged to provide any contingency.)

Problems associated with the operation of the sampler- Sampler
enricher caused more delay in the MSRE test program than
did any other component. Considering the basic difficulty
cf its task, it performed rather well, but its operation
was always rather ticklish. This was to some extent un- Management
avoidable, but every effort should be made to develop reactor
equipment that is robust and if not "foolproof'" is as
simple to learn to operate as possible. Anticipate foul-
ups, plan for maintenance, provide as much space as you
can afford: these are some more lessons from the sampler-
enricher,

The freeze valves are an example of a component whose General
cperation was something of an art that took some time to
master. The fewer items in this category the better from
the standpoint of startup delays, training requirements
and possible upsets during operation.

Emphasis on simplicity is goed, but it is possible Management
to overdo it. The reliability of the reactor was enhanced
in some respects by having only one nozzle on the reactor
vessel for three control rods and a sample array, by omitting
a flowmeter from the fuel circuit, and by drastically limiting
the number of thermocouple wells. It is questionable
whether the benefits outweighed the problems with the
flexible control rods, the embarrassing uncertainty in
the heat balance, and the limitaticns on the dynamic analyses
of the system.

 
Chemistry

Des - Comp & Sys.

Des - Misc.

Des - Instr.

Des - Comp & Sys.

Development

Construction

Organization

Planning

 

52

Reply M (Con't)

Some items fall into the category of failures to think
of everything in the design and specifications. Noteble
examples were the activation of the original corrosion
inhibitor in the treated water, the production of radio-
lytic gas in the thermal shield, and the condensation of
moisture in the component cooling air cooler, all of which
could have been anticipated.

One can't anticipate everything and a design should
have some flexibility in the form of spare connections,
redundant lines, and open space for additions. I think
the MSRE designers generally scored well in this regard.

The control and safety systems were carefully thought
out and designed. The MSRE I&C steering committee spent
many long hours with the future operators, exhaustively
(exhaustingly?) discussing possible situations and the
circuitry necessary to assure safe, relatively "tidy" operation.
It paid off in a reliable safety system and convenient
control. Of course most emergencies never arose and some
controls were never used except in test. An example is
the automatic load control system, which was elegant but,
as it turned out, an unnecessary frill.

Perhaps the biggest job of on-site development and
redesign was that which occupied all summer in 1968: the
fluorine disposal system in the processing plant. The
original design was regarded as proven and its operability
taken for granted because a similar S0,-F, reactor had
been operated elsewhere. The differences in operating
conditions and required efficiencies were not adequately
considered. When the deficiencies were recognized, an intensive
effort was mounted that led to a completely different system.
The development tests and redesign should have been done
earlier, preferably in another facility, to avoid the expensive
delay in the reactor schedule.

Construction

Construction of the MSRE was done with skill and great
care. Although the formal trappings were not as great as
are required by RDT today, the control was very good and
the result was evident in the cperation.

The overlapping of preliminary testing by the operators
and the final stages of construction is fraught with possibilities
for trouble. It takes a cooperative, understanding attitude
by both parties, clear definition of responsibilities and
careful planning., Given these factors, such a startup can
work well, and did at the MSRE.

 
53

Reply M (Con't)
Operation and Maintenance

Most of my notes on this phase of the MSRE seem to
be in the nature of items that I feel are worth emulating.
I confess to some difficulty in trying to be objective,
but I think the picture is not too slanted.

The operation and maintenance of the MSRE was always Personnel
in good hands. Management properly considered the MSRE
to be very important and made available a sufficient number
of qualified personnel from whom the MSRE staff was selected.
Nearly all of the supervisors had been involved in the
design and development effort and this was a very good
thing. It enabled the operating group to write training
manuals and operating procedures that were complete and
accurate.

Training of operators, preparation of operating procedures Training
and establishment of a filing system must be done very
thoroughly and carefully if a novel system (or complex
experiment) such as the MSRE is to be started up without
costly mistakes and is to yield the most information.
This was done at the MSRE.

For a novel system, the only way to do the best job Training
of preoperational training is by use of a simulator. MSRE
experience bore out the wvalue of this.

The MSRP and ORNL management showed their interest Management
and helped the MSRE staff and supporting forces to feel
the importance of their task. As a result, they dedicated
themselves to meking the reactor run well and were unstinting
in their efforts to get it back into operation quickly
when it went down. Although there was a sense of urgency
that speeded our efforts, I never felt pressured from
above to take shortcuts that the other members of the
staff and I felt to be unsafe or unwise.

Almost without exception, the people involved in Organization
the MSRE operation and maintenance were strongly mission-
oriented and red tape or organizational lines did not
hinder cooperation and participation.

I can't recall any decisions that were bad because
they were based on poor communications. In this regard, Communications
the routine planning meetings were very helpful. (These
were one-hour meetings starting at 11 a.m., one to three
times a week depending on the status of the reactor.) All
the available MSRE engineers met, usually together with
representatives of other parts of the MSRP, to hear a

 
Communication

Management

Maintenance

Changes

General

Computer

54
Reply M (Con't)

review of recent events, to report on their own work and

to discuss plans. As a result everyone was kept informed
and contributed to steering the test program. These meetings
were well worth the time,

In retrospect, I suppose it might have been better
to have replaced the fuel-pump rotary element at an early
date to stop the o0il leakage. But maybe not. We generally
heeded the mechanic's axiom: "If it ain't broke, don't
fix it!" which might be amplified to "if it is running well
enough, you may get into more trouble than you get out of
by trying to fix it." That still seems to be valid. This
is not to say that preventive maintenance is not necessary
or desirable. This was done systematically on the MSRE and
it paid off.

Expeditious maintenance and minimum shutdown times
require efficient planning and control. We had that on
the MSRE to a commendable degree.

Even in a system as well engineered and well-built
as the MSRE, lots of changes will be necessary. These should
be carefully reviewed, approved and controlled. The MSRE
Change Request system worked well, handling 761 change requests
between June 1965 and December 1969.

Some problems can be temporarily circumvented if the
physical system has enough redundancy or flexibility and
the people have enough motivation, ingenuity and thorough
knowledge of the system to devise remedies. A good example
of this at the MSRE was the alternate route for the fuel
offgas used in Run 10 to hold down particle trap temperatures.

The on-line computer at the MSRE was invaluable. No
reactor experiment should be without one. Planning and
programming should start early. It took much more time to
get the MSRE computer installed and debugged than we imagined
it would, partly because we were asking it to do things which
hadn't been tried before (even though its designers had
planned the capability).

 
55

Reply N

Construction, Inspection, Checkout, and Testing

Reactor Division construction and inspection group
was understaffed. There should have been a larger field
force working with the construction group.

The preoperational and precritical testing was very
important. Example: Accurate heater element resistances
were made using a Wheatstone bridge during preoperational
testing. A recheck of the resistances made after completion
of the precritical testing showed most of the heaters
were good, yet there was a slight increase in the resistance
of one fuel-pump heater. An investigation revealed a high
resistance contact between the heater element and the
extension lead. The weakness in the heater design was
corrected in all five removable fuel-pump heater units,
thus avoiding a difficult remote maintenance job after
power operation had started.

Design

More consideration should have been given to the
preheating and remote maintenance of the heaters on the
drain and fill line when the piping was designed.

The horizontal and vertical section of the fill and
drain line at the junction of line 103 and 104 was on
the same heater control. Although the heating was adequate,
it was time-consuming to operate within the temperature
limits.

MSRE heater control was manual because of the many
different heater configurations and wattages. In future
molten salt reactors, automatic program heatup and automatic
control for normal operation should be considered. This
should be possible with solid states devices.

Where possible, in-cell electrical connections should
be welded instead of screwed or bolted.

A metallurgical prcblem may exist if alloy-99 nickel
wire is operated at high temperature. There were several
extension lead wire failures in the radiator.

Construction

General

Des - Elec.

Des - Elec,

Des - Instr.

Des - Elec.
Des - Elec.
Des - Elec.
Training

Training

Operations

Computer

56

Reply O

Operator Training

Operator training was very thorough as were the pericdic
refresher courses which, incidentally, were needed because
of the long shutdown period for fuel reprocessing and 233y
addition; alsc during extended runs, there were few shutdowns
and restarts resulting in few opportunities for the cperator
not only to take the reactor critical but also to make the
transition from Flux to Temperature Servo mode. Although
testing is a necessary adjunct to any training program, it
was the least popular of the training phases.

The jumper board was a very good training and an even
better operating aid. Another aid to cperating personnel
was the posting of up-to-date flowsheets and engineering
drawings in various locations of the reactor building (such
as offgas flowsheet in the vent house, complete heater and
thermocouple drawings at the Heater Control Panel, cover-
gas flowsheet in the diesel house, etc.).

Computer Operations

 

The on-site computer was a dandy when it came to the
acquisition, comparison, calculation, and recording of data
and in other ways following the progress of reactor operations.
One of its functions, the alarm system, was somewhat troublescme
to operations personnel especially during transitions such
as drains, cooldowns, heatups, and fills. The essential alarms
for the reactor were on the Main Board. Quite a few of these
signals were also fed to the computer whose alarm point was
set somewhat below that for the Main or Auxiliary Board alarms
with the idea that the computer would give advance warning
that a variable was approaching the alarm point. Because
the incoming signal fluctuated or the internals of the computer
pickup were too sensitive, the computer would alarm, clear,
and realarm almost continucusly when signals approached the
alarm point. This also happened when some variables approached
zero or when a group of signals went out of limits due to
faults internal to the computer. In these cases, the audible
alarms from the computer were bypassed and subsequent type-
cuts on the alarm typewriter were sometimes ignored because
of continuous type-outs. On occasions, computer alarms resulting
from important calculated values which did not have duplicate
main board alarms were ignored or overlooked among the plethora
of other alarms that were typed out. In the future it would
probably be preferable to let the computer handle the entire
alarm system or if a dual system is used, there should be
no redundancy and the computer would then be restricted to
calculational alarms which would have more credence.

 
o7

Reply O (Con't)

There was a marked increase in interest in the computer
and its functions after some of the operators became "involved"
with the FOCAL program. Acceptance of the computer by
the cperating crew would probably have come socner, had
this program been initiated sooner.

Sampler-Enricher

 

The sampler-enricher is a complicated piece of machinery Sampler
considering the wvarious interlocks which had to be satisfied
in order to adhere to the stringent requirement of always Organization

maintaining double containment both during sampling and
transfer to the Hot Lab. Actual sample isolation from

the pump bowl was fairly simple using the written procedure; Operations
however, the transfer of the sample from one compartment

to another within the sampler and eventually into the shipping
cask required careful operator attention and a fair amount

of dexterity with the manipulators. Although training

sessions were conducted with a prototype manipulator at

the Y-12 plant and with the sampler during the intensive
sampling and enriching of Run PC-2, the interval between
consecutive samples later was too long to maintain proficiency.
For instance, on a rotating shift basis, weekly sampling

would afford each operator the opportunity to operate the
sampler only once every 4 months. The idea of having the
sample isolated on the midnight shift and having it transferred
and delivered by the regular day shift operators (who have
become more familiar with the sampler and procedure) was

a good one.

Although at first considered by the operating crew
to be awkward to operate and limited in its tasks, the
sampler-enricher proved to be far more durable and versatile Sampler
than expected. A total of TU5 sampling cycles including
152 enrichments was completed with essentially no contamination
problems except during repairs. Among the minor nuisances
were (1) buffer seal leaks at the access port, and upper
seal leaks at the operational, maintenance and removal valves,
(2) seventeen replacements of the flexible containment
membrane between the manipulator and main containment box.

Future design effort should be directed toward a better
seal at the access port and a more easily replaceable or
remotely reparable "1C" section.

A heavier membrane along with both a positive and
a negative pressure differential switch between Area 3A
and the manipulator cover would possibly reduce the rupture
frequency of the boot. Also, better illumination of the
1C area 1is needed.

 
Operations

Communication

Chemistry

Des - Ccmp & Sys.

Des - Comp & Sys.

58

Reply O (Con't)

Heater Control Panel

 

The idea of a "yellow" and "red" line setting for the
heater variacs and ammeters is a good one. The yellow-
line setting represented the variac or ammeter setting
which would maintain that section of the reactor piping
at 1100 - 1200°F when the reactor was empty. This was the
desired piping temperature before a fill and was close to
the operating temperature. During subsequent heatups,
the entire system could be heated more uniformly by speci-
fying a percentage of the yellow-line setting in a given
time interval. Also if the heater setting required a change
during operations, it could be reset to the yellow-line
value after a drain without resulting in a hot or cold
spot relative to adjacent piping.

The red-line setting represented a setting which was
not to be exceeded under any circumstances (mechanical stops
were fabricated for some variacs). A thermocouple on the
wall of a pipe containing circulating salt does not respond
as readily to a heater change as an enmpty pipe; thus the
temperature of a heater element could be raised to the
point of burnout while under the impression that the element
temperature is only & few degrees above the pipe wall temperature.
It is believed that heater life was prolonged by these
restrictions.

Offgas System

A great deal has been learned sbout fission gas and
decay daughter behavior in the MSRE and major obstacles
have been overcome with the offgas plugging problems; however,
fission product transport in the offgas stream is still
not completely understocod. The particle trap in line-
522 was very efficient at removing solid fission products
from the gas stream. There is a good chance that a second
particle trap at the exit of the second hold-up volume would
have trapped the decay daughters born in this section and
would thus have eliminated the plugging problem at the
entrance to the charcoal bed.

The plugging at the charcoal bed entrance was no doubt
hastened by the fact that the 1/2-inch Sched-L40 pipe
entered the charcoal bed flush up against the stainless
steel wool packing at the bed entrance section. Plugging
would probably not have been as rapid had the gas been
directed into the empty space above the wad of stainless
steel wool; the effective filtering area would have been
increased by a factor of 20.

 
59
Reply O (Con't)

Check Valves

Check valve failures have been a minor nuisance at Des - Misc.
the MSRE. Some of the offgas line check valves have developed
restrictions and some have come apart internally such as
those in the exit lines of the drain tank and in the discharge
lines of the component coolant pumps.

Freeze Valves
Early in the checkout of the reactor, the control Des - Instr.
circuits for the freeze valves were changed from a one-
out-of-two to a two-out-of-three circuitry. The associated
audible alarms were also changed such that no alarm occurred
when only one shoulder temperature exceeded its limit.
On frequent occasions when the heater, control module,
or control air setting either drifted or was improperly
set after a fill, the valve would be cycled from BLAST-
to-0OFF or from BLAST-to-HOLD-to-BLAST, etc. Had an audible
alarm been associated with the latter case, corrective
action by the operator would have been prompt thus fewer
thermal cycles would have been imposed on the freeze valves
(the flashing green light indicating BLAST AIR, with no
audible alarm, was often overloocked or ignored). Also
this signal could have been fed to the computer which Computer
could have kept tabs on the number of thermal cycles that
vere imposed on any freeze valve.

The freeze valve control modules (Electro Systems Des - Instr.
ET-4300) with their seemingly inconsistent alarm indications
were probably the least understood pieces of instrumentation
on the reactor. Also the necessity of manually clearing
the pot temperature modules (ET-4200) to get a thaw permissive
was often overlooked (this group should have been on automatic
reset).

 

Reply P
Chemical Processing Plant
(1) Corrosion, although lower than expected, was Des - Comp & Sys.
still one of the major problems. Future fluorinators should
either have frozen walls or use an alloy not containing Development

molybdenum. Absorbed MoFg prevented accurate determination
of the amount of uranium absorbed on NaF by weighing and
because of its high heat capacity, interfered with accurate

measuring of the UFg¢ in gas streams by means of the mass

 
Des - Comp & Sys.

Des - Elec.

Development

Computer

Training

60
Reply P (Con't)

flowmeters. A large gas filter was reguired downstream

of the caustic scrubber to prevent line plugging by molybdenum
compounds. The salt filter was effective in removing the
reduced corrosion products from the salt. However, most

of the metals remained behind in the processing tank and

will cause considerable difficulty in any future processing.

(2) Fibrous metal filters are effective in removing
metallic fission products {such as °°Nb) from gas streams.
All process gas lines leaving shielded areas should have
g filter at the cell wall.

(3) More attention needs to be given to design of
salt line penetrations through cell walls to prevent plugging.

(4) The KOH-KI solution provided a very efficient
means for removing excess fluorine from the offgas stream.
However, plugging of the dip lines needs to be prevented
either by laboratory work or equipment design.

Reply Q

General Comments

 

The application of a process control digital computer
in the MSRE for data acquisition was the first such application
at ORNL. In addition to the functions of monitoring the
reactor and recording data for further analysis, the system
provided valuable experience in real time digital data acquisition
and process techniques.

The one factor that perhaps could have made the computer
application more successful would have been to have more
of the reactor engineers mcre closely associated with the
computer system. During the life of the MSRE, the computer
project lost the services of two Mathematics Division programmers
(one on the BR-3L0 and one programming data retrieval of
the IBM-360) and two I&C Division engineers. In forthcoming
reactor projects involving the use of digital computers.
Reactor Division personnel (preferably engineers) should
be assigned to a computer application group during the
early phases of the project. They should attend the computer
manufacturer's programming course and should participate
in the initial planning for and programming of the computer.
It is these people who should be depended upon for con-
tinued support of the computer system through the life of
the project.

 
61

Reply Q (Con't)

Checkout and Acceptance

 

The system acceptance requirement of seven days' Computer
continuous operation without failure was a good one and
should be insisted upon for future systems. It took about Procurement
four months to get the system in condition to pass this
requirement. Due to the proximity of the computer to General

the reactor control room, the reactor operators witnessed
all of the "growing pains' of the system and consequently
were very wary of it when it was finally made available
for their use. Many months of satisfactory operation
were required to gain the confidence of the operators.

In future installations, a means should be provided
to "insulate'" the reactor operators from the computer
system until the system is completely checked out and
accepted. This may require only the installation of window
shades on any windows in the computer room and the postponement
of installing computer equipment (typewriters, alarms,
etc.) in the control room until the system is 99.9% ready.
The window shades would also be helpful for routine maintenance.

Maintenance

Experience with the computer seller's maintenance Computer
engineers was both good and bad depending on the experience
and personality of the individual engineer. When we eventually
purchased the computer, we sent two men to maintenance Maintenance
school and assumed system maintenance in-house. This
left something to be desired due to the following:

1. The ORNL maintenance men were not as familiar
with the computer as were the computer manufacturer's
maintenance men who had several years experience on the
particular model computer.

2. The computer manufacturer's documentation was Documentation
not adequate to support timely maintenance by anyone not
thoroughly familiar with the documentation. (The majority
of the documentation was not covered in the maintenance
course. )

3. A good knowledge of the computer programming
system is essential to effective maintenance. This knowledge
was lacking in both our and the computer seller's maintenance
people.

To remedy the above or future similar applications
we should:

1. Send at least two "hard core" employees to maintenance Training
and programming school as early as possible in the project.

 
Management

Maintenance

Computer

62

Reply Q (Con't)

2. Have these people assist in the initial programming
of the system and insist on their participation in the
factory acceptance tests,

3. Contract maintenance from the computer seller
for the first year of the project and have our people
work with the contract people on routine maintenance problems.

Hardware

The BR-340 computer system proved to be quite adequate
for the particular application. In fact, the computer
could have done much more than was required of it. The
digital computer state-of-the-art has advanced quite rapidly
since the acquisition of the BR-340, and the functions
it performed for the MSRE could probably be performed
now by two or three smaller computers at a substantial
reduction in cost. This would have the advantage of not
putting all of one's eggs in one basket. A failure of
a single computer would not result in a loss of all data
and computer functions while the one computer is out of
service. ©Such a multi-computer system should be considered
for future projects.

Although punched card readers and line printers are
expensive computer peripheral devices, the MSRE could
have benefitted considerably from the use of both. The
many hours spent editing paper tapes could have been reduced
to minutes 1f card input equipment had been available.
A line printer could have saved a considerable amount
of money that was spent on data searching and listing
at the ORNL computing center, the programming for which
was also expensive. Additionally, the printed data would
have been available to the analysis group much sooner.

Since there were only two magnetic tape units in the
computer system, some magnetic tape functions (such as
copying from one tape to another) were not possible without
losing reactor data. To circumvent this problem, the job
of tape copying and consclidation was transferred to the
ORNL computing center, but this proved to be quite a headache.
Future computer systems for applications such as that at
the MSRE should include at least three and preferably four
magnetic tape units. Here, again, the state-of-the-art
has progressed to the point where four tape units can now
be purchased for less than the price of the two units
on the BR-340. Also, when considering systems in the
future, attention should be directed to the magnetic tape
format of a computer word. The BR-340 can write a tape
that it can read and that can be read on another system,

 
63

Reply Q (Con't)

but it cannot read tape written by another system unless Computer
it is written in a special format peculiar to the BR-340.

The BR-340 writes {(or reads) 28 bits from five, six bit

frames of tape, low order bits first. Other machines

write 30 bits in five, six bit frames of tape and read

into or write the high order bits of a memory word first.

Software

The computer system programs supplied by the computer Computer
seller were generally very gocd. The standard type programs
and those written specifically for the MSRE by the computer
company's programmers required only minor modifications
during the term of the project. BSeveral man~years of ORNL
programming were required to update operational programs
(reactivity balance, heat balance, etc.) to current levels
of reactor operating experience. It is recommended that

cn future projects, Reactor Division engineers be trained Training
to do this type of program maintenance. Nuclear engineers
familiar with reactor technology, would be in a much better Organization

position to recommend additional functions for the computer
than I&C systems engineers.

Computer diagnostic programs were provided by the
computer manufacturer. These tested the components in
the system considered to be standard products by the manufacturer.
Unfortunately the MSRE analog signal input system was specially
designed for the MSRE and no diagnostic program was supplied.
Adequate hardware diagnostic programs must be required
of and demcnstrated by the computer manufacturer of the
next such system.

Reply R

I am pleased when my criticisms can be complimentary.
This is indeed the case for the operation of the MSRE.
Two points probably merit comment.

First, the system of marking and identifying each Qual Assurance
piece of material that went into the MSRE was very gocod.
Most projects start out incorporating such a program, but
the bookkeeping becomes less complete as the project proceeds.
I know of no other proJject in which the steps can be retraced
from any part in the Reactor to a vendor's certification
sheet giving the chemistry and heat treatment of this material.
This has been very valuable to use in our materials work
and a similar system should be followed in future reactors.

 
Des - Comp & Sys.

Management

Organization

Procurement

6L

Reply R (Con't)

Second, the two surveillance facilities that were provided
were in excellent location for following the behavior of
the materials used in the Reactor. OQur own reople, namely
Howard Cook and Al Taboada, deserve some credit for their
design of the surveillance assemblies that were used. The
attitude of the MSRE staff toward the surveillance program
was very good, with removals and reinsertions being handled
with high priority and care.

Reply S

It is my belief that the delegation of well-defined
administrative and technical responsibilities ought to be
made in order to obtain maximum efficiency in carrying
out future reactor experimental programs. The delegations
need not be exactly as suggested but they ought to be fitted
to the talents of available key people. Tt is intended,
however, to stress that the individual to whom a specific
responsibility is delegated by the Project Director also
be assigned administrative control of the associated work
and the resources to be used. It has been my experience
that when the requirement for using persuasion to obtain
maximum contribution from members of a cooperative system
is substituted completely for administrative control, the
system breaks down and becomes less efficient and less
effective.

Reply T

Comments and Observations on Fabrication of the MSRE Major
Nuclear Components

 

The following comments and observations on fabrication
of the major nuclear components for the MSRE are offered
in response to the memo, MSR-T0-6, by Paul N. Haubenreich.

Introduction

The MSRE major nuclear components were fabricated
from Hastelloy-N (INOR-8), a new alloy.

Private industrial companies were contacted in 1961
about fabrication of these components. However, they were
unable to bid because of their lack of eXperience with
this alloy and the need to develop new fabrication methods,

 
65

Reply T (Con't)

techniques, and procedures. Experience in fabrication Procurement
of Hastelloy-N, prior to the MSRE was very limited. It

had consisted of fabrication in Y-12 and ORNL Shops of

small parts for test loops.

Because of the additional development work needed Management
for fabrication and to obtain the high quality desired,
it was decided by MSRE project management to fabricate
the major nuclear components in UCC ND Shops where close Construction
liaison could be maintained between design, development,
and the shops. With the exception of the graphite core
for the reactor and selected parts, the reactor vessel,
primary heat exchanger, both salt pumps, four salt storage
tanks and the salt-to-air radiator were fabricated and
assembled in the Y-12 General Machine Shop.

The successful operation of these components at the
MSRE site indicate that the high quality fabrication deemed Qual Assurance
necessary and the resultant policies, methods and procedures
used in the fabrication were adequate and justified. Many
of the policies, methods and procedures that were used
were ploneering, especially in the field of quality assurance.

Management

Shop management was informed early in the MSRE program Management
of the need for successful operation of the major nuclear
components in the MSRE if the molten salt reactor concept
was to survive. They understood the need and cooperated
ccmpletely by furnishing their best manpower and equipment
for all phases of the job. The shop assigned a master
craftsman to each component who worked on that component Organization
exclusively from the time fabrication was started until
it was completed. This master craftsman was specially
selected for his skill, dedication, pride and leadership.
He assisted the shop foreman in coordinating the work of
all supporting craftsmen.

Critical path scheduling was successfully used for Management
the first time in ORNL for scheduling the fabrication
of the major nuclear components. This was a joint effort
between shop management and the Fabrication and Construction
Group of the MSRE project. Shop management assisted in
planning the step-by-step operations and estimated the
time required for each operation. This participation
by shop management presented them with the opportunity
tc plan the entire job in advance, select manpower and
tools, and recommend design changes for ease of fabricatiocn. Changes
Many of these changes were made provided they did not
affect quality or performance. Of course, the main benefit

 
Organization

Communication

General

Communication

66

Reply T (Con't)

was to give an estimated completion date based on the best
information available at the time and to indicate the critical
operations that required attention to prevent time delays.

An effective organization was established to forward
information back and forth between the MSRE project and
the shops. ©OSee Figure 1. The MSRE project selected a Shop
Coordinator, who was experienced in shop practices, to work
directly with shop and inspection persconnel. He furnished
the shop with all information and special materials including
Hastelloy~N. In addition he was responsible for cbtaining
answers to questions that originated in the shop during
fabrication. The Shop Coordinator worked directly with
the Project Technical Coordinator. The Technical Coordinator,
an engineer, originated all purchase orders, requests for
material, and requests for fabrication bases on design drawings,
specifications, and other project needs. The Technical
Coordinator was also responsible for resolving all shop
originated problems and, when necessary, after consultation
with the appropriate Project design group, metallurgist
or other specialists.

Regularly scheduled meetings, usually once a week,
were held between shop management and MSRE project personnel

to exchange information and discuss problems.

Esprit de corps

 

Enthusiasm, devotion to duty and pride in good workmanship
by shop management, and especially the craftsmen, were outstanding.
It is difficult to assess how this motivation was accomplished,
however, it probably was due to the fact that shop management
and shop craftsmen were included as participants in a team
effort with the MSRE project to fulfill the need of developing
a successful Molten Salt Reactor. This participation included
the following:

(1) Seminars and discussions were conducted by Project
personnel to give background information on the MSRE and
to inform shop management and craftsmen of the project
needs.

(2) Careful attention was given by the Project and
shop management to planning by critical path scheduling
and attention to detail so that materials, information,
manpower and equipment were available at the proper time
or with minimum delay.

 
 

6T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

- h . t
2 Metals & Ceramics ) g gzggiists
; (A. Taboada) l p
]
Technical Coordinator
(C. K. McGlothlan) _;
'
1
; Shop Coordinator
i (J. W. Teague) i
e S
!
|
. r—
[ Shop Management Inspection ? Audit
| L L

 

 

 

 

Figure 1. Information Flow for Fabrication of MSRE Major Nuclear Components
Qual Assurance

Documentation

68

Reply T (Con't)

(3) Project personnel were respectful of the skills
and talents of each craftsman. Personal contact by Project
personnel was maintained with craftsmen concerning status
and problems of their work. Craftsmen were frequently photo-
graphed with the component they were fabricating. Personal
copies of these photographs were given to each craftsman.
This small gesture was significant in creating pride in
good workmanship.

(L) Problems and questions originating in the shop
were always resolved as quickly as possible by Project Coordi-
nators. This action was considered very important by the
Project in order to maintain esprit de corp and efficiency
in the shop.

Quality Assurance

To assure that the MSRE major nuclear components would
be fabricated to the highest nuclear standards and fulfill
project needs a quality assurance program was developed.

It was a pioneering effort in this field and provided an
integrated plan utilizing new and some existing quality
assurance policies, procedures and methods. Some of the
more significant elements of the quality assurance program
were:

(1) All drawings, specifications, and instructions
were reviewed by the Technical Coordinator for material,
welding and other fabrication requirements consistent with
the appropriate quality levels established by the Project.

(2) Strict material specifications were used to purchase
Hastelloy-N. The weldability test requirement in the specifications
was extremely important because slight variations in standard
industrial air melting techniques could adversely affect
the welding characteristics of this new alloy. The measurement
and documentation of all chemical and physical properties
to assure that all material met minimum requirements for
strength, ductility, etc., was also an important requirement
of the material specifications. Only one quality, reactor
quality, Hastelloy-N material was accepted and received
by UCC-ND. Material inspection was usually performed utilizing
the vendor's inspectors and equipment with auditing by
qualified inspectors from the ORNL Inspection Engineering
Department,

 
 

69
Reply T (Con't)

(3) An integrated Quality Assurance Program Plan' Qual Assurance
was developed and used for all fabrication. It included
procedures for material control, special fabrication, cleaning,
heat treating, assembly, inspection, testing, shipping,
and record keeping.

(4) Special Hastelloy-N fabrication procedures were
developed and used for the first time. These procedures
included welding and weld inspection, brazing, braze joint
inspection and assembly of the primary heat exchanger.

A1l welding was performed by welders that had been trained
and qualified by the Y-12 Weld Test Shop.

(5) Fabrication activities were documented. They Records
included results of cleaning, heat treating, weld inspection,
pressure and leak testing, and material traceability for
all Hastelloy-N parts and components from raw stock through
all fabrication operations.

(6) Qualified inspectors from the Y-12 Mechanical Qual Assurance
Inspection Department were used to provide independent
(independent of shop management) inspection services for
verification of material identity, weld inspection, pressure
testing and auditing of shop dimensicnal and liquid penetrant
inspection activities. Auditing of the Y-12 Mechanical
Inspection Department's activities was performed by qualified
inspectors from the ORNL Inspection Engineering Department.

Directing the fabrication of the MSRE major nuclear
components was a challenging and satisfying job. It was
a pleasure to work with so many capable and dedicated
people. This wealth of capability and dedication, sometimes
overlooked, is an asset that should not be allowed to dissipate.

 

1
C. K. MeGlothlan, Fabrication Specificaticns, Procedures,
and Records for MSRE Components, MSR-62-3, Rev. A, (Feb. 1965),

Internal Memo for Limited Distribution.
Des - Instr.

Management

Management

 

70

Reply U

The following comments and observations are submitted
for possible inclusion in your compilation of criticisms
of the MSRE design. This critique is submitted with the
realization that some design and management deficiencies
cited were fully or partially remedied dquring the course
of MSRE design and operations and that others will be reme-
died by better future planning and by advancement in the
state-of-the-art.

General

Although the performance of the MSRE was not perfect
and undoubtedly could be significantly improved if we had
a chance to design it again, I believe it was generally
satisfactory and beyond expectations. Where design defici-
encies existed, I believe they were mainly due to one or
more of the following factors:

(a) The development nature of the project.

(b) The necessity for fitting the system into an existing
facility (Bldg. 7503).

(c) A tight and unrealistic initial schedule.

(d) Lack of firm criteria and recognized standards
for design of containment systems.

(e) Excessive "economizing" in the area of safety
instrumentation.

(f) The "state-of-the-art" of solid state and data
acquisition equipment in the early 1960's before the MSRE
design was frozen.

Scme of these factors are interrelated and others were
possibly unavoidable at the time. Certainly (a) and (f)
and possibly (b) were unavoidable but it is less apparent
that (c), (d), and (e) could not have been avoided by more
detailed planning during the initial phases of the Title
I design. Of these, the most painful and possibly the
most Important was the tight schedule which resulted in
a heavy overlap in Title I and Title II design and required
the use of much "judicious guessing" in the areas of process
design as well as instrumentation and controls. In retro-
spect, I am surprised that everything came together as well
as 1t did and that more scrapping of design was not required.
I believe that the Title I design should have been completed
and thoroughly reviewed before the Title II design was started.
T1

Reply U (Con't)

In particular, I believe that process flowsheets should
have been completed before detail design of the process
piping and I&C systems was started. As it was, the process
flowsheets were not available until after the instrument
application drawings were nearly complete and these were
generated by extracting information from variocus designers
as the system grew.

The lack of firm criteria and recognized standards
for the design of safety system resulted in much confusion,
and many heated discussions before agreements were reached.
The net result was a complicated, patchwork, design, which
while accepted, really did not satisfy anyone. This was
particularly true in the area of containment and process
safety instrumentation. Design philosophy and criteria
for nuclear safety instrumentation, though not documented,
was much better established and problems encountered in
this area were not severe. Had we had a comparable philosophy
for containment and process safeties (or better yet an
accepted standard) the complexity of the MSRE safety systems
could have been reduced and we would have had a more relisble
system.

In addition to having accepted criteria and project
design standard, I strongly recommend that all designers
be made aware of the existence and need for them. Awareness
of these requirements would, hopefully, result in the
designers taking the lead by building the needed containment
and/or redundancy into the design instead of having two
teams — one which designs a system to meet functional reqguire-
ments and another which tries to plug all the holes in
the system containment with instrumented closures. For
example, much of the containment safety instrumentation
could have been eliminated if the treated water system
and portions of the cover-gas system had been built inside
secondary containment. Similar examples exist in other
parts of the MSRE system.

Data Acquisition

 

(A) Computer-Data Logger — The Computer-Data Logger
proved to be a useful tool. Future reactors should be
provided with similar equipment and more of the routine
monitor and control should be delegated to it.

 

(B} Temperature Scanner — Although the performance
of this system was not as good as the operators would
have like, it was better than its designers expected. I
believe it served its intended purpose of providing a relative
low cost method of monitoring large quantities of temperature

Des - Instr.

Management

Management

Procurement

Computer

Des - Instr.

 
Des - Instr.

Des - Instr.

Des - Instr.

Management

Des - Instr.

T2

Reply U (Con't)

data. The most troublesome part of this system was the
mercury Jjet commutator switches. Similar systems with
solid state switches are now available at reasonable costs
and should be considered for use in future reactor systems.

(C) Safety Systems — There should have been more extensive
use cof two-out-of-three coincidence logic in the safety
systems. One-of-two logic is safe but it does not permit
testing.

There should have been more separation of safety and
control circuits. The use of valves and other final elements
for both control and safety resulted in complexity of control
circuits, difficulty in maintaining separation of safety
and control wiring and, in my opinion, a less reliable
safety system than we could have had. I realize that there
will probably always be areas where it will be impossible
to use separate device for safety and that control and
providing separate devices is (initially) expensive; however,
I believe that a stronger effort to obtain this objective
in the MSRE would have paid dividends in savings in design
time, and in increased system operability and safety system
reliability.

In retrospect, the use of undercurrent relays in the
"Galloping Safety" circuits to detect loss of fuel flow
conditions was a mistake. The basic idea of implying loss
of fuel flow from reduction of pump current was basically
O.K. but was difficult to implement under the constant pump
speed conditions when all failure modes were considered
and completely fell apart when the pump was operated at
reduced speed. Use of power instead of current trips would
have been more direct, less troublesome, and more compatible
with variation of pump speed. This approach was initially
considered but was abandoned for reasons of economy and
expedience.

(D) Fuel Salt Flow — A greater effort should have
been made to find an acceptable method of measuring fuel
salt flow. If a fuel flowmeter had been available, the
problem discussed in C above would never have occurred
and the data obtained would have been valuable for other
purposes. (I have never been convinced that the venturi-
NaK transmitter type flowmeter used in the cooclant system
would not have been acceptable for fuel flow measurement
and consider the decision not to use it as arbitrary.)

(E)} Stack Flow — If a venturi or flow nozzle had been
installed on the stack as originally proposed, a good heat
balance across the radiator could have been obtained and
the uncertainty about the reactor power might have been
avoided.

 
T3

Reply U {(Con't)

(F) Annunciators — The annunciator system performed
acceptably considering the number of alarm points and
the state-of-the-art when the MSRE was designed. However,
in the future, I think the smaller, solid state, units
now available should be used and that more of the alarm
monitoring function should be delegated to a Computer-
Data Logger. ©Solid state logic modules should also be
used to replace the "Rochester" units.

(G) Temperature Switches — Temperature alarm switches
used in future systems should have high input impedance
so that they would be insensitive to variation of the
resistance of lead wire and connection points in the input
circuits and so that the calibration would not be dependent
on the value of these resistances. Also, cold juncticn
compensation should be provided when these switches are
actuated by thermocouples.

 

(H) Valves — An effort should be made toc develop
valves for offgas system service which are less susceptible
to plugging.

(I) Valve Operators — The effort spent developing
the motion multiplier for the long stroke valves located
in the reactor cell would have better been spent on develcoping
a long-stroke, radiation-resistant valve operator. Such
an operator will undoubtedly be needed in future reactors
and none is available.

 

(J) Pressure Transmitters — If absolute instead of
gage pressure transmitters had been used for measurement
of gas pressures in the contained systems, the problem
of providing secondary containment for these transmitters
would have been much simpler.

(K) Penetration Seals — Providing seals on the sheathed,

 

multi-conductor, glass-insulated, thermocouple wire penetrating

the reactor vessel proved to be more difficult than had
been anticipated. Strong consideration should be given
to using sheathed, multi-conductor, mineral insulated
cable for similar applications in the future.

(L) Electrical Noise in Signal Wiring — The use of
standard, unshielded, thermocouple lead wire was a mistake,
In order to provide adequate electrical noise reduction
(and improved accuracy) all thermoccouple wire as well as
other low level signal wiring should be twisted and shielded.

 

Des -~ Instr.

Computer

Des -~ Instr.

Development
Development
Des - Instr.
Des - Instr.
Des = Instr.
Des - Instr.

 
Des - Instr.

General

Maintenance

Remote Maint.

Remote Maint.

Procurement

Th

Reply U (Con't)

(M) Interconnection Wiring Design — Design of the
contreol circuit interconnection wiring consumed an inordinate
amount of time. Most of this time was spent on "bookkeeping"
operations of the type that could be easily performed by
a computer if the techniques and programs were worked out
in advance. However, I do not suggest that all control
circuit wiring could be turned over to a computer. Panel
wiring design involves many arbitrary decisions which can
be better done by people.

 

Reply V

The comments listed below pertain to the MSRE. A
comparison to the comments made of the HRE in 11/7/52
(CF-52-11-139) indicate many improvements have been made
but note the following comments from the above memo: "Building
too small. Access for maintenance and inspection should
be improved. Adequate working room around equipment. Specify
uniform bolt and nut size in a limited number, etc." It
would appear that we do not take full advantage of past
experience in following through on many valid comments
regarding our own machines.

Programmed Maintenance

 

Careful preparation of the routine computerized programmed
maintenance functions based on manufacturer's information,
working knowledge, and awareness of the MSRE system requirements
were responsible for the success of this program. Operations
group preparations and approval, maintenance work performed,
and follow-up procedures functioned well and routinely.

Building 7503 Facility

 

(a) The head room available in the high bay was insufficient
for the length of the tools required for much of the in-
cell remcte maintenance work. Clumsy and difficult manual
manipulation of the long-handled tools was required due
to interference with the crane bridge.

(b) The high-bay ventilation system created an air
turbulence which contributed to the spread of radiocactive
contamination throughout the high bay. Decontamination
of this area was expensive and time-consuming, especially
during periods of maintenance activity.

 
5

Reply V (Cen't)

(¢) The only means of removing any object from either
the reactor or drain cell was by pulling the object into
the high bay after placing the object in a suitable container.
One instauce required removal of a valve (HCV-523) from
the reactor cell into a shielded carrier loccated in the
high bay without first being placed in a cecntainer. A
sealed, shielded work or storage area having direct access
to the reactor cell and accessible from another location
would reduce the radiation and contamination hazard to
maintenance personnel.

The primary hot maintenance tool, the portable maintenance

shield, was in constant use during remote maintenance operations.

Suggested improvements:

1. better lighting,

2. more Viewing ports,
3. more work plugs,
4,  improved bearing surfaces, and

55 dual, independent rotating work plugs.

There were more than ten sizes of bolts used in the
reactor cell assembly, each of which required a different
tool head. Regardless of the bolt size, this number could
be standardized to three or four bolt sizes and reduce
tool costs. In some locations, special machined thread
bolts were used when standard good gquality rolled threads
would have been stronger and served the same purpocse.

The 522 line contained an unsupported flange (the
outboard jumper line flange) which became twisted and
out of level when torque was applied during leak-testing.
This caused difficulty in remote assembly and the threat
of twisting off the 1/2-in. pipe below the flange when
torquing.

Additional spare thermocouple disconnects and power
leads located within the cell would have been useful.

Maintenance required at the sampler-enricher was frequent
and complicated by the contamination and radiation hazard.
The most common failures were the urethane boots which
formed the seal around the manipulator arm. The boots
were changed 18 times at a cost of ~ $2000 per change.
The chief reason for the failures appeared to be that the
work area in the S/E was too small or the manipulator

Remote Maint.

Remote Maint.

Management

Des~- Mech

Des -~ Mech.

Remote Maint.

Des - Mech.

Remote Maint.

Des - Instr.

Sampler

 
Sampler

Des -~ Misc.

Des - Misc.

Remote Maint.

Des - Misc,

Des - Comp & Sys.

Development

76

Reply V (Con't)

too large to permit free use of the arm in either the extended
or retracted positicn. A second manipulator could have

been used to good advantage. Blind spots existed in the

work area.

An on-site decontamination area would have been useful
on many occasions.

The vent house — offgas area was mass of organized
confusion due to space limitations. A modular hot cell
arrangement would have simplified the required piping
and eased the maintenance problems.

The water room was too small for the amount of equipment
installed, the net result was a labyrinth of overlayed

piping.

Reply W

The criticism which is presented here has to do with
the problem of plugging in the pump tank offgas circuits
of the salt pumps during operation of the MSRE. This problem
is apparently the result of two conditions: a possible
small leakage of oil from the lower shaft seal oil leakage
catch basin into the pump tank and the accumulation of
aerosols which are carried out of the pump tank with the
purge gas. Had these conditions been recognized early
in the development of the salt pumps, solutions might
have been developed or at least their contribution to the
problem might have been reduced. At least provisions
could have been provided to minimize the effects,

A solution to the oil leakage problem was developed
by providing a seal weld between the shield plug and the
bearing housing. However, this solution was develcped
after the salt pumps were installed at the reactor. The
spare rotary elements were modified with this seal weld,
but the leakage problem was not deemed significant enough
to require replacement of the original rotary elements.

It was late in the development of the pumps that the
salt aerosol problem was detected. Previously, the high
flow rate (b 2/m) of offgas required at the MSRE had not
been tested. Therefore, solutions and schemes for coping
with this problem were provided as modifications at the
MSRE.

 
1

Reply X

While there were weak points in the design, mistakes Management
in extrapolating the development models into the actual
system, and errors made in the operation of the MSRE, these
were so few in number that it is hard to be critical of
any aspect of the program. I am sure that all associated
with the program agree that the decisions made by thoese
in managerial positions were superb and were so often proven
to be excellent decisions that we of the "Monday-morning
quarterback club" had little to discuss.

I do have this suggestion. The training program in Training
which the operations engineers and technicians participated
was thorough and well administered; however, in retrospect,
I feel it was deficient 1In one important aspect. There
were no instructions for the engineers in the principles
of supervision and/or management. These may be learned
with experience, but they may be learned with much less
pain in the classroom. I would suggest that for future
cperations in which technical people are going tc be thrust
into the role of supervisors that some management training
be given them, particularly to those that are inexperienced.

Reply Y

In general, MSRE worked so well that I am unable Des - Misc.
to make any really constructive suggestions beyond the
obvious ones, all relatively mincr. In a broad sense I
guess all I can say is that during the design we got into
detailed design too socon, a "hurry up and wait' philosophy Management
which caused delays, which may have been avcidable, after
we were presumably ready to operate. Possible remedy —
"Forbid hiring detail draftsmen until a good basic design
is well laid out and confidence therein has been established."

Again, since it was such a successful operation,
this is in the category of nit-picking, hindsight, perhaps.

 

Reply Z
MSRE Better Points
1. I feel that the fuel and coolant circulating Des - Comp & Sys.
pumps were the best pleces of mechanical equipment in the
Jjob.
2. The piping design also was very good. We spent Des - Piping

locts of time on the big reactor Jig in the crane bay. 1
sometimes wonder if all the precision measurements were

 
Des - Elec.

Remcte Maint.

Des - Comp & Sys.

Des - Misc.

Des - Misc.

Management

Management

Des - Misc.

T8

Reply 7 (Con't)

necessary. Could we have done the piping in reactor cell
then had engineers record all the dimensions and elevations,
etc.?

3. 1 feel the remote maintenance shield is a real
good piece of engineering, but the next one needs to be
improved. Chiefly, the maneuverability of it, such as
installing a track on two sides of the cell so it can be
meved similar to a bridge or overhead crane.

L. I think much thought should be given to C.C.P.
installation. I would suggest a cell or compartment with
bulk-head doors that are easily sealed and unsealed and
that the cell be large enough to perform the required
maintenance on the equipment inside it. This, of course,
is to eliminate all the physical moving of shielding and
unbolting and rebolting of containment vessels. This
could cut the down time also.

5. The next reactor should certainly have a contaminated
work area properly designed adjacent to the reactor to
enable us to do maintenance and repair quickly and effectively.

6. Keep as many mechanical components (such as control
valves) as possible outside the reactor cell.

7. Above all — NEVER build another reactor in an
existing barn and well.

 

Reply AA

I believe that the project was fortunate in not having
any equipment or component failures that precluded the
success of the experiment.

One area in which an improvement over procedures used
for the MSRE could be made is in the design approval process.
I think that a formal design review committee should have
been established to approve not only the design drawings
(as was done), but also the calculations on which they
were based. That is, calculations supporting the design
would accompany the drawings when submitted for approval;
the design review committee would be responsible for verifying
the correctness of both.

Granted that design and development reports were issued
on major and important items, these generally appeared
after the equipment had been built. The purpcse of having
a formal review of design calculations would be to provide

 
 

T9

Reply AA (Con't)

an independent check in order to eliminate errors such

as those made in the radiator heat exchange calculations,
and to insure that future molten salt reactor projects
are as successful as the MSRE.

Reply BB

Some Comments Relative to the Cover Gas and Offgas Systems

 

Cover Gas System

 

Des - Comp & Sys.
In light of the excellent corrosion experience in
the fuel salt system, we may want to consider relaxing
the concentration limits for water and oxygen in the helium Management
cover gas. Originally, we were not sure what level was
tolerable. Average analyses for U. S. Bureau of Mines
helium showed less than 1 ppm for water and oxygen, so
1 ppm was selected as a reasonable value. The small purification
system was added as a precautionary measure. To the casual
observer, the 1 ppm seems conservative. Thus for MSRE
conditions of 4.2 liters/min helium and about 10,000-
1b salt, l-ppm H,0 v/v in the helium would produce about
0.008 ppm/day U0, w/w in the salt, and it would take about
34 yr for a 100-ppm rise in UO, concentration. This is
great except that 1 ppm is at the lower end of the range
for a standard moisture analyzer and the problem of routine
monitoring would be a lot simpler if the allowable concentration
could be raised to 3 or 5 ppm.

In considering the design of the system, it might Des - Piping
be argued that we could have gotten by with cheaper materials,
e.g., copper tube instead of stainless steel pipe, packed Materials

instead of bellows-sealed valves, etc. In my opinion,

the added confidence and reliability which was obtained

by using the higher quality materials was well worth the
incremental cost, and the same philosophy would be even
more applicable in the case of a 30-yr power reactor system.

Offgas sttem

There seems to be a feeling in some quarters that Des - Comp & Sys.
the charcoal beds were over-designed because they were
able to operate with only one section in service without
any noticeable ill effects. This could be an illusion. Management
The beds were designed for T.5-days Kr (equivalent to
90-days Xe) holdup. The holdup test, which we ran during
the prepower period, indicated that the actual holdup was
about 6-days Kr (72-days Xe), which is only 80% of the
intended value. When operating with only one section,

 
Des - Comp & Sys.

Des - Instr.

Management

Des - Piping

Materials

80

Reply BB (Con't)

the holdup time would be 1/2 or 36 days for Xe. The attached
table shows calculated activities for the longest-lived

isotopes, 12 d '*'MYe and 5.27 d '*%Xe, in the offgas stream

at the charcocal bed outlet for 36-days Xe holdup. The figures
show that the activity should be very low. For purposes

of comparison, it might be noted that the values are lower

than those which are permissible for continucus exposure

to the general public (see NBS Handbock 69). It may be

that the reason the beds seemed to be doing so well is that

the activity of the gas was near or below the minimum sensitivity
of the detecting instruments. During the post-mortem,

we should re-run the Kr holdup test to determine the effect

of 100,000 Mwhr of operation on the hcoldup time, and samples

of charcoal shculd be removed (after a sufficient decay

time) and analyzed to determine the fission product distribution.

Needless to say, we failed to anticipate and make
proper provision for the plugging problem. A replaceable
mist trap and filter should have been provided at the pump
bowl outlet. A filter was installed at the control valve,
but it was inadequate for the job and had tc be supplanted
by the so-called particle trap, which apparently was quite
effective. The design at the inlet to the charcoal bed
should have been plug-proof and possibly should have included
a replaceable filter.

We should have provided more and better diagnostic
tools. Thus, we could have had more thermocouples on the
offgas line to get a better picture of the temperature profile.
We could have provided a means for obtaining samples of
the offgas at the pump bowl outlet. We subsequently provided
a gas sampler at the outlet end of the in-cell volume holdup,
but it came late in the game and was bothered by plugging
problems. We should have taken gas samples at the charcoal
bed outlet. This would have shed some light on the question
of charcoal bed performance.

The remarks made regarding quality of construction
in the cover-gas system may be repeated with even greater
emphasis for the offgas system. To my knowledge, there
was no case of an operating interruption due to a leakage
of radicactive material from the offgas system. This is
an argument for the use of quality materials and also is
a tribute to the skill of the men who put the system together.
A bad leak in the offgas system would have been costly if
not disastrous. :

 
 

81

Calculated Activity of Longest-Lived Isotopes
in

MSRE Offgas Stream at Charcoal Bed Outlet

 

 

 

U-235 Activity of Gas
Half Constant Xe Holdup Time Fission Yield Ci
Life A th Y As Micro
i t-1/2 (sec™!) (sec) (%) (dis/sec+cc) (Curies/cc)
131mye 124 .668 x 10-° 3.11 x 10° 0.03 1,500 0.0bL
133%e 5,27 1.52 x 10-° 3.11 x 10° 6.5 52,500 1.4

 

At 1 MW(t) and assuming t 1/2 is long relative to the reactor residence
time:

Rate leaving reactor = 3 x 10t Yi atoms/sec; where Yi = fission

 

yield in %.
Let: Power = 8 MW(t)
Flow = L4200 cc/min = TO ce/sec
Ai = decay constant, sec™' for isotope i
th = holdup time, sec
A, = activity of isotope i, dis/(sec)(cc)
Ci = Curies/cc for isotope, 1
Then:
A, = 05 1018 (v,)(2;) o~ ty)
h.2 x 10°
A, = 60 x10%° (YD) &M tn) ais/(see) (co)
And:
c, = 16 (Y )(A) o~ (A ty) Curies/cc
82

APPENDIX I

A Request for Criticism of the MSRE

 

A letter with this title was written by P. N. Haubenreich on
January 16, 1970 and was distributed as MSR-70-6 to 76 persons who had
been associated with the MSRE. The text of the letter was as follows.

"What have you learned from the MSRE? What avoidable de-
ficiencies have you observed in its design, construction,
operation, maintenance or analysis? On the positive side,
what good points are worth repeating?

"You are receiving this letter because we believe that through
your assoclation with the MSRE you probably have learned some
lessons that are worth sharing or emphasizing as we turn

from the MSRE to the tasks that lie ahead of us. In an

effort to insure that none of our hard-won lessons are lost,
we plan to publish a critique of the MSRE, We solicit your
contribution.

"What we have in mind is something similar to the critique
published by Sam Beall (ORNL-CF-61-6-116) at the end of
the HRT coperation. The scope of cur critique will be as
broad as you wish to make it, but in order to be of most
value, each comment should be as explicit as possible. The
critique will be widely read (we hope) within the project,
and as always tact is desirable; but do be candid.

"I have asked Ralph Guymon to receive and categorize your
contributions, to crganize and publish the critique. Your
comments and suggesticns will not be censcred; any editing
that seems desirable will be cleared with you. Each of
your remarks will be identified with your name unless you
request that any or all be anonymous.

"Please give this critique your serious attention and
submit your contribution to R. H. Guymon, Building 7503,
as soon as practical. If you are unable to participate,
please let him know."

 
were submitted by the following persons.

request for criticism (Appendix I) except for

83

APPENDIX II

List of Contributors

The 28 contributions which are reproduced

which were submitted (to S. E. Beall) when he

Reply

A

£ =M |3 g O

= HOR Y49 o

W 9d 9" o O Y 49 49 46 = HH I

-

Name

. Blumberg

G. Bohlmann
L. Compere

. Cook
. Crowley
. Engel
Franzreb

. Fry
. Kryter

Gabbard
. Gallaher
. Grindell

. Guymon
. Harley

H od =B @ o @b o= X)X 9 B&om

. Haubenreich

in the body of this report

All were in response to the

W. B. McDonald's comments,

left ORNL in 1965.

=
0
E

td
WENMNZ‘<C.'H(D':UQ"UOZ

> X o= &G W o= oW o a = o @ 9ok 13

B R &0 B O @ H B

= o YU 3" o a @

Name

Hudson

Krakoviak

. Lindsuer

Martin

McCoy

. McDonald
. McGlothlan
. Moore

. Richardson

Smith

. Steffy

Tallackson
Todd

. Ulrich

Smith

 
8L

APPENDIX ITI

Tepical Index

The following topics are discussed in this report. An underlined
number is the page where the comments on this topic are summarized.
Other numbers are the pages where the topic is mentioned in individual
contributions.

Margin Notations

 

Changes - 11, 13, 5k, 65
Chemistry - 6, 10, 20, 21, 29, 32, 47, 51, 52, 58

Computer - 7, 11 , 13, 14, 25, 27, 46, sk, 56, 59, 60, 61, 62, 63,
71, 73

Construction - 6, T, 29, 4k, 52, 55, 65
Containment - 7, 43, L8

Design (Components and Systems) - 11, 12, 13, 1k, 15, 22, 28, 30, 31,
32, 3T, 48, k9, 51, 52, 58 59 60 6h 76 TT 78 79, 80

Design (Electrical) - 8, 12, 2k, 30, 33, Lk, 55, 60, 78

Design (Instrumentation) - 7, 8, 12, 13, 1L, 15, 23, 24, 26, 27, 29,
33, 34, 35, 43, Lk, 47, h8 hg, 52 55, 59, TO L, T2,
73, T4, 75, 80

Design {(Mechanical) - 8, 10, 12, 1k, 18, 28, 34, 36, 75
» 8, 12, 14, 17, 25, 29, 34, 35, 45, 50,

 

Design (Miscellaneous) - 6,

52, 59, 76, 77, T
Design (Piping) - 11, 12, 1L, 29, 30, 48, 49, 77, 79, 80
Development - 6, 7, 11, 12, 30, 32, 50, 51, 52, 59, 60, 73, 76

T
8

Documentation, Records, and Communication - 6, 9, 13, 14, 16, 18, 19,

24, 35, 36, 37, 39, ko, b1, L2, 43, L6, hT "S53, 5&, 58 61,
66, 68, 69

General - 6, 13, 49, 50, 51, 54, 55, 61, 66, Th
Maintenance - 6, 7, 10, 13, 23, W7, 5Lk, 61, 62, Th
Management and Planning - 6, 7, 8, 9, 11, 13, 1k, 17, 18, 19, 20, 21,

22, 23, 26, 28, 29, 33, 3h 35 _56 38 39, ho he h3 hh
us L6, L1, 51, 52, 53, sk, 62, 64, 65, T0, T1, 72, 75, TT,

78, 795 80
Materials of Construction - 13, 39, 79, 80
Operations - 6, 7, 9, 22, 25, 28, 30, 35, 37, L4, 45, 50, 56, 57, 58

Organization and Personnel - 6, 13, 16, 17, 22, 23, 25, 37, 4k, 52, 53,
57, 63, 64, 65, 66

 
85

Margin Notations

 

(continued)

Procurement - 7, 34, 38, 39, 48, L9, 61, 6L, 65, 71, T4

Quality Assurance - T, 13, 22, 29, 63, 65, 68, 69

Reactor Analysis - 6, 11, 26, 27, 33, 3k

Remote Maintenance - 10, 1k, 16, 17, 18, Tk, 75, 76, T8

Sampler - 7, 10, 20, 21, 25, 28, 36, 37, 47, 51, 57, 75, 76
Training - 9, 11, 13, 17, 23, 35, 45, L6, 53, 56, 60, 61, 63, TT

 
11,

12.

13.
1L,
15.
16.
17,
18.
19.
20.
21,
22,
23,
ok,
25,
26.
27.
28.
29.
30.
3l.
32,
33.
3k,
35.
36.
37.
38.
39.
Lo,
L1,
Lo,

L.

O o3 O\ W

MSRP Director's Office
. Affel

. Anderson

. Baes

. Ball

R.
J.
C.
S.
H.
S.
M.
E.
D,
F.
R.
E.
C.
G.
R.
S.
D.
J.
R.
C.
E.
W.
D.
L.
J.
F.
R.
J.
S.
F.
W.
J.
E.
D.
L.
A.
J.
D.
J.
A.
C.

HHAH A Q@

Beuman

. Beall
Bender

S. Bettis

S. Billington
F. Blankenship
Blumberg

G. Bohlmann
J. Borkowski
E. Boyd

B. Briggs
Cantor

HdoRmE Y EEdoderpdaptt3ddnpe s o =

Cardwell
Chandler
Chepman
Collins

. Compere
. Cook

Cope (AEC-OSR)
Corbin
Crowley

. Culler
. Debakker
. Distefanc

Ditto
Doss

. Eatherly
. Engel

. Epler

. Ferguson
. Ferris

. Fraas

Franzreb
Fry

Frye
Fuller
Gabbard

87

DISTRIBUTION

oo @omYRdxNvsaadarydanmPlEddnd 3 i S GG dd e 5

e

oDy adg=Eonmaxyu

g rHEHOSTHOsS a3 44 m »H

AEHPPOUHGOEHK

ORNL-CF~-T0-9-3

Gallaher

. Grimes

Grindell
Guymon

. Harley

Heubenreich
Helms
Herndon

. Hightower
. Hill
. Hise

Hoffman

. Holz
outzeel

. Hudson
. Jackson
. Jordan

Keplan
Kasten
Kedl
Kelley
Kerlin
Kern
Keyes
Kirslis
Koger
Korsmeyer
Krakoviak

. Krewson

Kryter
Lamb
Lane

it Laughon, AEC-0OSR
. Lindauer

. Lundin

. Lyon

. MacPherson

MacPherson
Martin

Matthews, AEC-0OSR
McCoy

. McGlothlan

 
99.
100.
101.
102.
103.
10k,
105.
106.
107.
108.
109.
110.
111.
112,
113.
11k,
115.
116.

RO aoatbmernSE PR R

T
H
L
J
H
A
C
R
H
E.
L.
H
A
H
B
G
J
M
R
H

. McLain

. McNeese

. McWherter
Metz
Meyer
Mills
Moore
Nichol
Nicholson
. Cakes

. Payne

. Perry

. Piper

. Prince

. Ragan
Redford

. Richardson

C. Robertson

. M. Roth, AEC-ORO

Dunlap Scott

H.
J.

E. Seagren
H. Shaffer

McIntosh, AEC-

88
ORNL=-CF=T70-9-3

Distribution

(continued)

Washington 117. M. J. Skinner
118. W. L. Smalley, AEC-ORO
119. A. N. Smith
120. W. F. Opencer
121. I. Spiewak
122. R. C. Steffy
123. R. D. Stulting
124, J. R. Tallackson
125. W. Terry
126. R. E. Thoma
127. W. D. Todd
128. D. B. Trauger
129. R. W. Tucker
130. W. C. Ulrich
131. W. E. Unger
132. G. M. Watson
133. B. H. Webster
134. A. M. Weinberg
135. J. R. Weir
136. M. E. Whatley
137. J. C. White
138. Gale Young
139. F. C. Zapp

M. Shaw, AEC-Washington

140-141,
142,
143-145,
146.
147,

Central Research Library
Document Reference Section
Laboratory Records

Laboratory Records, Record Copy
ORNL Patent Office