Keywords: EPRI NP-3111

EPRI
Electric Power
Temporary Repair
Leak Sealing
Sealant Extrusion
Project 1328-1
Final Report
May 1983
Research Institute Sealant Chemistry
Mechanical Sealing

EPRI-NP—3111

DEBS 902677

Testing and Evaluation of

On-Line Leak Sealing Methods

Prepared by
Combustion Engineering, Inc.
Windsor, Connecticut
 DISCLAIMER

This report was prepared as an account of work sponsored by an

agency of the United States Government. Neither the United States
Government nor any agency thereof, nor any of their employees,
makes any warranty, express or implied, or assumes any legal liability
or responsibility for the accuracy, completeness, or usefulness of any
information, apparatus, product, or process disclosed, or represents
that its use would not infringe privately owned rights. Reference
herein to any specific commercial product, process, or service by
trade name, trademark, manufacturer, or otherwise does not
necessarily constitute or imply its endorsement, recommendation, or
favoring by the United States Government or any agency thereof. The
views and opinions of authors expressed herein do not necessarily
state or reflect those of the United States Government or any agency
thereof.

D IS C L A IM E R

Portions of this document may be illegible in electronic image

products. Images are produced from the best available
original document.
Testing and Evaluation of On-Line
Leak Sealing Methods

N P-3111
Research Project 1328-1

Final Report, May 1983

Prepared by

COMBUSTION ENGINEERING, INC.

Nuclear Power Systems
1000 Prospect Hill Road
Windsor, Connecticut 06095

Principal Investigator
G. A. Smith

Project Manager
N. R. Stolzenberg

Prepared for

Electric Power Research Institute

3412 Hillview Avenue
Palo Alto, California 94304

EPRI Project Manager

B. P. Brooks

System Performance Program

Nuclear Power Division MASTER
RSTOBUnOM OFTHIS DOCUMENT IS UfU
 ORDERING INFORMATION
Requests for copies of this report should be directed to Research Reports Center
(RRC), Box 50490, Palo Alto, CA 94303, (415) 965-4081. There is no charge for reports
requested by EPRI member utilities and affiliates, U.S. utility associations, U.S. government
agencies (federal, state, and local), media, and foreign organizations with which EPRI has an
information exchange agreement. On request, RRC will send a catalog of EPRI reports.

NOTICE
This report was prepared by the organization(s) named below as an account of work sponsored by the Electric
Power Research Institute, Inc. (EPRI). Neither EPRI, members of EPRI, the organization(s) named below, nor any
person acting on behalf of any of them: (a) makes any warranty, express or implied, with respect to the use of
any information, apparatus, method, or process disclosed in this report or that such use may not infringe private­
ly owned rights; or (b) assumes any liabilities with respect to the use of, or for damages resulting from the use
of, any information, apparatus, method, or process disclosed in this report,

Prepared by
Combustion Engineering, Inc.
Windsor; Connecticut
 EPRI PERSPECTIVE

PROJECT DESCRIPTION

Temporary on-line leak sealing of components (such as valves, flanges, and pipes)
that is performed while the system remains in operation may provide the plant oper­
ator with an alternate to an unscheduled plant shutdown to effect a permanent
repair. This temporary repair may be cost effective in avoiding the loss of revenue
resulting from the shutdown and additionally may prevent damage to the component
(due to steam cutting, etc), which requires extensive and costly maintenance to
achieve a permanent repair.

This project (RP1328-1) evaluates the effectiveness of commercial leak repair ser­
vices, with particular emphasis in the application of the on-1ine sealing techniques
in nuclear power plants. No attempt is made to determine if one vendor, furnishing
this leak repair service, is superior to another. The work performed in the project
is intended to provide guidelines to assist utilities in making the decision on
whether to apply this on-line leak sealing method and what to expect if the decision
is made to apply it.

PROJECT OBJECTIVES

The objectives of this project are to determine the limitations of the use of on­
line sealants to temporarily eliminate leaks in process systems in nuclear power
plants and to provide conclusions and recommendations on the use of these materials
and techniques.

PROJECT RESULTS

Three companies, which provide on-line leak sealing services, participated in the
project by demonstrating their equipment and techniques used in sealing a variety of
leaks. A fourth company participated in a limited part of the total test program.
The leak repair equipment and techniques used were commercially available at the

iii
time the tests were conducted (August 1980 to January 1982). The participation of
the following companies in this project is gratefully acknowledged:

• Pipe Line Development Company (Cleveland, Ohio)

• Federal Industrial Services, Inc. (Redford, Michigan)

• Leak Repairs, Inc. (Alvin, Texas)

• Furmanite (Virginia Beach, Virginia)

The above order of participants does not necessarily correspond with the identifica­
tion of Vendors A, B, C, and D shown in the text of the report. No single organiza­
tion was clearly superior to the others.

Performance results indicate that leak sealing of valves was generally less success­
ful than leak sealing of pipes and pipe flanges. A number of recommendations were
developed as a result of the work performed. Utilities contemplating the use of an
on-line leak sealant service should consider, in detail, each of the recommendations
noted in this report and should work closely with the organization that will perform
the service.

Each of the above listed participants was given the opportunity to review and to
comment on this report. One of the participants provided a 1ist of items that
should be evaluated by potential users of on-line leak sealing services and re­
quested that the 1ist be included as an addendum to the report. This, however, was
not incorporated because the list generally parallels the conclusions and recommen­
dations of the report. Briefly summarized, potential users are encouraged to com­
prehensively evaluate leak repair services in areas such as, but not 1imited to, the
following:

i Methods used to prevent injection of sealants into the process

stream

• Effects of the sealant materials, fixtures, and leak repair methods

on the system pressure containment boundary

• Qualification and applied controls in the application of the repair

This report should be useful as a guide to maintenance personnel and to those in

engineering and management who will be making the decision to apply these on-line
leak sealing materials and techniques. The majority of this project was managed by

iv
T. W. Libs and R. E. Swanson. However, questions on the content of this report
should be directed to B. P. Brooks.

B. P. Brooks, Project Manager

Nuclear Power Division

v
 ABSTRACT

A test program was conducted to evaluate the effectiveness of commercial

leak repair services, with respect to the requirements of nuclear power
plants. Three companies commercially engaged in on-line leak sealing
demonstrated their sealants, fixtures and techniques on an evaluation matrix
of leaking pipes, flanges and valves. Repaired test specimens were left
on-line at full internal fluid pressure and temperature of approximately
2200 psi and 600°F (15.2mPa, 316°C) initially for 48 hours. Selected

specimens were left on-line at pressure and temperature with thermal and
pressure cycling for as long as 14 weeks to determine longer-term repair
durability. Sealant samples were thermally aged, irradiated, and analyzed
for the presence of reactive components such as sulfur, heavy metals and
halogens.

Test results and observations showed that the effectiveness of an

injected-sealant repair is determined by the quality of the initial repair,
and by its performance within the first 48 hours following the repair. Test
results indicate that sealant extrusion into the process stream can be
inhibited but not positively prevented in all test cases. Thermal eye 1ing
of otherwise effective repairs frequently causes deterioration and premature
failure. Repairs of valve bonnet flange and packing gland leaks were
generally less successful than those of leaking pipes and pipe f1anges.
Principal recommendations based on the tests are that users: a) consider
these repairs only for short-term use; b) consider the added mass of repair
fixtures with regard to seismic requirements for piping systems; c)
obtain a certified analysis of the seal ants used to detect the presence of
reactive components; and d) evaluate the effects of the highly pressurized
sealant and mechanical sealing operations on the primary pressure boundary
structure.

vi i
 CONTENTS

Section Page

1 INTRODUCTION 1-1

2 OBJECTIVES 2-1

3 TEST PROCEDURES AND FACILITIES 3-1

Test Procedure 3-1

Test Facilities 3-2

4 SEALANT EXTRUSION TEST 4-1

Vendor A 4-1

Vendor B 4-2

Vendor C 4-3

5 ON LINE SEALING TECHNIQUES 5-1

Vendor A 5-1

Vendor B 5-3

Vendor C 5-4

6 TEST RESULTS AND DISCUSSION 6-1

Effectiveness of Repairs 6-1

ix
 CONTENTS (continued)

Section Page

Relative Advantagesand Disadvantages 6-2

Vendor A 6-2

Vendor B 6-3

Vendor C 6-4

7 MECHANICAL SEALING TEST 7-1

Vendor A 7-1

Vendor B 7-2

Vendor D 7-3

Summary and Conclusions 7-4

8 CONCLUSIONS AND RECOMMENDATIONS 8-1

Conclusions 8-1

Recommendations 8-3

9 REFERENCES 9-1

APPENDIX A VENDOR SURVEY A-l

APPENDIX B SEALANT MATERIAL ANALYSES B-l

x
 ILLUSTRATIONS

Figure JLeIS.6

4-1 Original Sealant Extrusion Test Fixture 4-5

4-2 Modified Extrusion Test Fixture 4-6

5-1 Vendor A Enclosure on 4 in Flange 5-7

5-2 Comparison of 3/4 in Flange Repair Fixtures 5-8

5-3 Vendor B 2 in Valve Repair 5-9

5-4 Vendor C Flange Repair Showing Injection Sequence 5-10

5- 5 Comparison of 3/4 in Line Enclosures 5-11

6- 1 Extrusion of Sealant Into Body of 2 in Valve Sealed by

Vendor A 6-9

6-2 Peening Damage on 4 in Flange Sealed by Vendor A 6-10

6-3 Close-Up of Peening Damage on 4 in Flange Sealed by

Vendor A 6-10

6-4 Sealant Extrusion Found in Pipe and Strainer Downstream

of 3/4 in Pipe Sealed by Vendor B 6-11

6-5 Leak in Vendor C Line Enclosure Prior to Termination

of Test 6-12
 ILLUSTRATIONS (continued)

Figure Page

6-6 Vendor C Line Enclosure with Extruded Sealant Slugs 6-13

6- 7 Section Through 3/4 in Pipe Sealed by Vendor C, Showing

Bulk Extrusion 6-14

7- 1 Vendor A Mechanical Sealing Fixture, Installed 7-5

7-2 Vendor A Mechanical Sealing Fixture, Disassembled After

Test 7-6

7-3 Vendor B Mechanical Sealing Fixture 7-7

7-4 Vendor B Mechanical Sealing Fixture, Disassembled

After Test 7-8

7-5 Flange Removed from Vendor B Mechanical Test Fixture 7-9

7-6 Vendor D Mechanical Repair Fixture, Installed 7-10

7-7 Vendor D Mechanical Repair Fixture, Disassembled After Test 7-11

B-l Test Sequence for Sealant Samples B-10

xi i
 TABLES

Table Page

1 Initial Leakage @ 48 hours 6-5

2 Results of Long-Term Thermal Cycling Tests: Vendor A 6-6

3 Results of Long-Term Thermal Cycling Tests: Vendor B 6-7

4 Results of Long-Term Thermal Cycling Tests: Vendor C 6-8

B-l Seal ant Material Analysis: Vendor A B-6

B-2 Seal ant Material Analysis: Vendor B B-7

B-3 Sealant Material Analysis: Vendor C B-8

B-4 Limits of Elemental Impurities in Non-Metal lie Materials

Established for Combustion Engineering Nuclear Steam Supply
Systems B-9

B-5 Sealing Performance of 3/4-inch Flanges During Oven Aging B-l 3

B-6 Sealant Hardness Measurements During Oven Aging B-14

xi i i
 SUMMARY

A test program was performed at Combustion Engineering, Windsor, Connecticut,

to evaluate commercial leak repair techniques that were available during the
time of these tests (August 1980 to January 1982). Three firms engaged in the
business of on-line industrial leak sealing participated in the test program:
Vendor A, Vendor B and Vendor C. Vendors were required to seal small hole
leaks in 3/4 in (19mm) pipes, 3/4, 2, and 4 in (19, 51 and 102mm) flanges, and
bonnet, stem and packing leaks in 3/4, 2 and 6 in valves (19, 51 and 152mm).
Repairs were made to the leaking components while under line conditions of
pressurized borated water at up to 600°F, and 2200 psi (316°C, 15.2 mPa). Sealed
specimens were left on-line for a minimum of 48 hours, after which they were
slowly cooled, depressurized, and checked for leakage and evidence of sealant
extrusion. Certain specimens were reinstalled on-line for long-term durability
tests of up to 14 weeks, while the remainder were set aside for material testing.
The long-term test specimens were subjected to thermal and pressure cycling
between atmospheric and 1ine conditions, and valves were periodically operated.

Test results were inconsistent. Each vendor had incidents of sealant

failures and extrusions, as well as 100% effective repairs. Valves in
particular proved to be difficult to seal effectively. Generally, repairs
which were effective after 48 hours survived the long-term testing. Thermal
cycling in particular contributed to the deterioration or premature failure
of the repairs.

Sealant behavior during injection was found to be unpredictable, with

frequent incidents of off-gassing and expulsion of sealant from the repair
fixtures. Si nee extrusion of seal ant into the process stream is a serious
concern for industrial plants contemplating the use of leak repair services,

S-l
it is noteworthy that no positive means of preventing such extrusion was
demonstrated. Further, the chemical composition of the seal ants and their
stability under long-term exposure to temperature and radiation varies. It
is in the best interest of the users to identify the presence of any
potentially reactive or hazardous components of the seal ants prior to their
use on-line, and additionally to require that an evaluation be made of the
effect of the high pressure seal ant injection process on the primary
pressure boundary structure of the component being sealed.

S-2
 Section 1

INTRODUCTION

The operational environment of components in a nuclear power plant causes

them to be subjected to high temperature and pressure with instances of
severe thermal cycling over the plant life. Valves, flanges, piping and
welded joints are critical parts of the system in controlling and containing
the coolant flow. Due to the high pressure and thermal cycling of the
system, leaks develop in such areas as valve packing or gaskets, welded
joints, and between mating flanges. If the magnitude of the leaks becomes
1arge enough, the pi ant may be forced to shut down prematurely in order to
make the necessary repairs. However, if such leaks can be sealed with
acceptable techniques and sealants while the pi ant is in service, operation
may continue until the next scheduled outage when repairs can be effected.

To determine the state of the art in the practice of sealing on-line leaks
with regard to line conditions at nuclear power plants, a test program
sponsored by EPRI was conducted at the test faci1ities of Combustion
Engineering, Inc., Windsor, Connecticut, during the period of August, 1980,
to January, 1982. Three firms actively and competitively engaged in on-line
industrial leak sealing participated in the major program - Vendor A, Vendor
B and Vendor C. In addition, a fourth vendor (Vendor D) was invited to
participate in an adjunct test of mechanical sealing fixtures. Vendor C
declined to participate in that particular test. This report documents the
objectives and performance of the test program, and draws conclusions
regarding leak sealing techniques and materials demonstrated by the
participants.

1-1
 Section 2

OBJECTIVES

The objective of the leak sealing test program was to have several
commercial vendors of the leak repair services demonstrate their products
and techniques by sealing a variety of test specimens (valves, flanges and
pipes) under different line conditions. These repairs would be evaluated in
the short and long term (up to four months) for cold leaks, seal ant
extrusion into the process stream, and ease of seal ant removal. Chemical
analyses would be used to determine the presence of potentially harmful
elements (e.g., halogens) as well as to define the stability of sealant
samples under thermal and irradiation aging. In addition, a number of
industrial users of leak sealing services was to be surveyed to gain
perspective on the performance of these repairs in the field.

2-1
 Section 3

TEST PROCEDURES AND FACILITIES

TEST PROCEDURE

Per Reference 1, each participating vendor was required to demonstrate his

sealing technique for the following on-line leak conditions using borated
water as the test medium:

Small Hole Leaks In 3/4 in (19mm) Schedule 160 Pipe

1/64 in (0.4mm) dia hole 0 540oF, 2200 psi (282°C, 15.2mPa)

1/32 in (0.8mm) dia hole @ 200°F, 2200 psi (93°C, 15.2mPa)
1/16 in (1.6mm) dia hole @ 600oF, 2200 psi (316°C, 15.2mPa)
1/16 in (1.6mm) dia hole @ 200°FS 2200 psi (930C, 15.2mPa)

Flange Leaks 1/32 in (0.8mm) Score Across Mating Flange Faces

3/4 in (19mm) flange @ 200°F, 2200 psi (93°C, 15.2mPa)

3/4 in (19mm) flange @ 600°F, 2200 psi (316°C, 15.2mPa)
4 in (102mm) flange 0 600°F, 2200 psi (316°C, 15.2mPa)

Valve Leaks

3/4 in (19mm) bonnet valve with 187 IX packing; packing leak @ 600°F, 2200
psi (316°C, 15.2mPa)

2 in (51mm) bonnet valve with 1871/grafoi1 packing; bonnet leak and packing
leak at 200oF, 2200 psi (93°C, 15.2mPa)

6 in (152mm) nuclear grade globe valve with 1871/grafoi! packing; stem

packing leak and bonnet seal leak @ 600°F, 2200 psi (316°C, 15.2mPa)

The test specimens are accurately described in References 2 through 8.

3-1
Once the specimens were sealed they remained on-line at the designated
temperature and pressure conditions for a minimum of 48 hours. The
specimens were then cooled and depressurized to atmospheric pressure, and
checked for sealant extrusion and leakage. Certain specimens were held
aside for material tests (refer to Appendix B), and the remainder were
reinstalled on-line for longer-term durability testing.

TEST FACILTIES

Tests were conducted on two hot loops of the Nuclear Laboratories at the C-E
Engineering Development and Services Department. Leak specimens sealed at
200°F and 600°F (93°C and 316°C)5 simulating primary flow conditions

in a pressurized water reactor, were installed in a dedicated test section

of the TF-2 large hot loop. The test medium in this facility was borated
water. This facility is capable of 15,000 gallons/minute (57 cubic
meters/minute) flow at 650°F and 2500 psi (343°C, 17.3 mPa). The

smaller TF-16 loop, capable of 100 gallons/minute (380 cubic

decimeters/minute) at the same temperature and pressure, was used for
simulating secondary side steam conditions and for certain long-term,
on-1ine tests.

3-2
 Section 4

SEALANT EXTRUSION TEST

The objective of the sealant extrusion test, which was to quantify and
relate sealant behavior to injection pressure and system temperature, was
not achieved. It was hypothesized that a family of temperature-dependent
curves relating injection pressure to cure time could be developed, which
would thereby determine a maximum safe injection rate for a given
temperature. This test was based on the belief that during the injection
process, the sealant develops a "cured front" which progresses to the
incident point of the leak. As the cured front obscures the leak, an
increase in injection pressure is noted as cured seal ant is forced into the
incidence to seal the leak.

Data obtained from the extrusion test failed to yield any meaningful
correlat ions among the parameters of interest, i.e., injection rate,
injection pressure, volume of seal ant injected, and temperature. However,
the following observations are presented which reflect the behavior of the
seal ants under the specific test conditions imposed.

VENDOR A

The seal ant extrusion test was performed by Vendor A on two separate
occasions. During the first test, attempts at injection of seal ant at
temperatures above 400°F (204°C) were made impossible by off-gassing of

the sealant material. The off-gassing condition would either cause a

portion of seal ant to be forcefully expelled through the orifice, or would
expand and separate the sealant into several chunks, thereby resulting in a

4-1
cured front in the tube itself, rather than at the orifice (See Figure
4-1). At temperatures of 300°F (150°C) or below, off-gassing was less

of a problem, but curing time increased dramatically such that a single test
exceeded 8 hours. Further, it was discovered that seal ant extrusion
continued until the injection valve was shut. No reasonable curing time
would prevent further extrusion if reinjection was attempted.

Discussion among C-E, EPRI and Vendor A following the first test led to a
retest. The original test fixture (Figure 4-1) was modified to provide back
pressure, and a small vent hole was located about 1 in (25 cm) from the
orifice (see Figure 4-2). Vendor A personnel performing the test were
allowed to pump by feel, altering injecting rates as their experience
dictated. Both problems of off-gassing and continuous seal ant extrusion
failed to materialize during the retest.

When questioned about the off-gassing of the sealant during the first test
and the complete absence of this condition during the retest. Vendor A
proposed that a difference in storage conditions may have been responsible
for the change in sealant behavior. The batch of sealant used in the retest
had been kept in deep-freeze storage prior to use, in accordance with Vendor
A's standard storage procedure. The sealant initially tested, however, had
been out of freezer storage for almost a month before its use. If improper
storage can result in poor sealant performance, then users should determine
specific storage requirements and assure vendor compliance.

VENDOR B

The requirements of the sealant extrusion test were incompatible with the
on-line sealing procedures employed by Vendor B. The extrusion test was
designed to relate a maximum allowable injection over-pressure to a given
injection rate. However, Vendor B by procedure does not exceed system
pressure during the injection process, thereby reducing the possibi1ity for

4-2
sealant extrusion. Despite this procedural incongruity, the Vendor B
sealant adapted well to the test conditions, with good results at 400°F
and 600°F (204°C and 316°C). The basic sealant cures very slowly at
temperatures below 700°F (371°C), so for this test Vendor B was allowed

to use an additive (claimed to be "nuclear grade") to speed up curing.

Injection pressures in excess of 2500 psi (17.2 mPa) (actual sealant
pressure) were obtained in under 15 minutes with no extrusion. Off-gassing
of the seal ant, while not uncommon, usually did not result in expulsion of
significant amounts of sealant through the orifice. A phenomenon observed
during the Vendor B test and associated with extrusion was a spontaneous
increase in pressure occurring 2 to 6 minutes after termination of
injection; however, this occurred only in tests with high (8 strokes/minute)
injection rates.

The Vendor B uncured sealant is fairly fluid, and cures to a cardboard or

asbestos-1 ike consistency. Uncured seal ant should be handled with gloves,
as direct skin contact has caused a severe rash on certain persons.

VENDOR C

The Vendor C seal ant compound proved to be incompatible with the conditions
of the extrusion test. At temperatures of 500°F (260°C) and above, the

seal ant would either off-gas, causing violent sealant expulsion, or

decompose, 1 osing its sealing properties. (According to Vendor C engineers
present during testing, this loss of sealing properties occurs when oxygen
is present and there is a lack of a steam environment.) At 400°F
(204°C), tests were time-consuming and inconsistent. Upon the formation

of a stringer, extrusion continued indefinitely, even when injection was

terminated and curing time allowed before reinjection. In one instance the
seal ant was allowed to cure for 25 minutes after an initial stringer was
produced with an injection tool pressure of 2400 psi (16.5 mPa) (not seal ant
pressure). After curing, extrusion resumed at the marginally higher
pressure of 2600 psi (17.9 mPa). By comparison, 1000-1500 psi (6.9-10.3
mPa) pressure is required to force uncured sealant from the injection gun
into the atmosphere. It should be noted that "injection pressures" reported
for Vendor C are hydraulic injection tool pressures, and not actual seal ant
pressures.

4-3
In its uncured form, the Vendor C sealant is a dense, rubber-1 ike material.
After short-term exposure (48 hours) to primary coolant conditions, bulk
sealant samples lost their surface resilience, and presumably any
thermo-plastic nature.

During handling of Vendor C's sealing compounds, a single individual

developed an apparent allergic skin reaction. Subsequent investigation
substantiated this belief. Discussion with Vendor C, while failing to
identify a seal ant constituent which could have caused this skin reaction,
did reveal the presence of bound acrylonitrile in the seal ant elastomer,
releases of which Vendor C reported do not exceed OSHA limits by
calculation. No similar information was provided by the other vendors;
however, prudence dictates that all pertinent precautions be taken in the
handling and use of any sealants.

4-4
 -.37^ <f> X .0<£>7

Figure 4-1. Original Sealant Extrusion Test Fixture

 Small Diaireter Hole
Orifice Simulating Nitrogen Supply
Unsealed Portion Of Leak To Be Sealed For Backpressure
Injection Gun
Tube Enclosure

-1^
I
CT>

Original Extrusion Collection

Test Fixture Chamber

Figure 4-2. Modified Extrusion Test Fixture

 Section 5

ON-LINE SEALING TECHNIQUES

VENDOR A

The sealing processes demonstrated by Vendor A were basically identical for

all specimens tested. The leak is first peened or otherwise mechanically
diffused, greatly reducing its magnitude. A two-piece sealant retainer, or
"clamp," is then installed. The cl amp has three separate cavities for
seal ant containment; one on either end of the longitudinal axis to form a
"perimeter seal," and one encompassing the leak, referred to as the "void".
Each cavity is equipped with multiple ports to allow for the injection of
sealant. The perimeter is sealed (injected) first and al1 owed to cure for
approximately 20 minutes before injecting the void.

The first injection into the void is made through a valved port
approximately 180° from the leak, while venting through a diametrically

opposed valve. When seal ant begins to extrude through the vent, that valve
is closed, and another valve is opened. The point of injection may or may
not be changed depending on the size of the cl amp and number of injection
ports provided. Ideally, as the void is nearly filled with seal ant, the
leak path will be reduced to two adjacent valves. At this point, if the
leak is sealed by injecting through one of these valves with the other
vented to atmosphere, the possibi1ity of introducing a stringer is minimal.
Occasionally, the leak wi11 become isolated to a single port; usually an
adjacent valve is opened and the cured seal ant drilled to reestablish a vent
prior to injection. If the vent path cannot be reestablished, the vendor
injects through the final valve, which increases the possibility of
introducing a stringer. A typical injection sequence of the void (refer to
Figure 5-1) is as follows:

5-1
 Injection Number Injection Port Vent

1 5 1
2 5 3
3 6 1
4 6 2
5 1 2
6 4 3
7 3 2
8 3 I
9 2 3

The initial leak in this case was between valves 2 and 3. Other valves are
utilized for providing the perimeter seal. Item 7 is the down stream
strainer used to determine if seal ant was carried into the process stream.
Reinjections of al1 or some of the ports after 1 hour is common.

Valve packing leaks are repaired by drilling and tapping into the stuffing
box, then injecting pumpable packing material until the leak is sealed and
ful1 travel is restored to the gland.

Rate and total volume are not monitored during the injection process. The
procedures demonstrated during the tests, however, call for a final
injection gun hydraulic pressure of 9000 psi (62 mPa). It is impossible to
relate this to injection or cavity pressure since the slightest restriction
in the flow path increases the hydraulic pressure dramatically. For
example, it takes 2000 to 3000 psi (13.8-20.7 mPa) pressure to push uncured
seal ant out of the gun by itself. Hydraulic pressures during initial
injection range from 4000 to 6000 psi (27.6 to 41.4 mPa). It should be
noted that these pressures refer to hydraulic injection tool pressures, and
not seal ant pressures.

5-2
VENDOR B

Vendor B demonstrated a variety of sealing techniques. Most of the flanges

were sealed with clamps, but with only one cavity at the void for sealant
containment. These clamps fit between the flange faces and around the bolt
circle O.D. The clamps were installed at full leak conditions with no
initial reduction of the leak by mechanical diffusion or other means. Once
the clamp is in place, the O.D. of one flange is peened at the clamp face,
restricting the leak to the flange studs and the clamp injection ports.
Since the sealant used is fairly fluid, the injection process takes
advantage of the pressure drop from the clamp interior to atmosphere. A
large portion of the sealant is blown out by the leak during the early
stages of injection; however, the leak eventually becomes smaller due to
surface coating and progressive sealant build up. When the leak becomes
quite small and isolated, the injection rate is decreased drastically (1
stroke/10-15 sec) until the leak is sealed. At the moment the leak stops,
system pressure equals line pressure and the injection valve is closed. The
gun remains attached for a short time before removing it from the injection
valve. Total volume and injection rate are not monitored; however,
injection pressure is not allowed by procedure to exceed system pressure at
any time during the process.

Because of the fluid nature of their sealant, Vendor B was able to use fewer
injection ports on their flange enclosures. The 4-inch flange required only
two injection ports, in contrast to Vendor A's six and Vendor C's eight .
Despite the installation of four injection ports on the 3/4 in (19mm) flange
enclosures (see Figure 5-2), the bulk of the sealant was injected through
the first port. After the first injection the selection of ports for
additional injections was random, but the final injection was always made
through the port nearest to the slowly diminishing leak, which typically was
situated at a stud.

5-3
If the pressure closures are designed such that the gasket is physically
retained at the pressure boundary, such as in the case of a captured gasket
or metal 0-ring, Vendor B may elect to inject directly into the damaged
gasket cavity by first drilling, tapping and installing an injection port
and valve in the area of leak. This was the method used for both 6 in and 2
in (150mm and 51mm) valve bonnet leaks. See Figure 5-3.

The clamps used for the 3/4 in (19mm) pipe specimens were unique in that
they did not require the injection of the sealant into the area surrounding
the leak. The seal ant is used as a gasket material between mating surfaces
of the two halves, while the longitudinal axis is sealed on either end by
means of a two-piece compression fitting which bolts around the pipe and
into the clamp.

VENDOR C

Vendor C flange repair fixtures (Figures 5-2 and 5-4) closely resembled
those used by Vendor B in that they fit between the flange faces, leaving
only the cavity surrounding the leak for seal ant injection. The injection
ports, however, are installed in the flange itself, rather than in the
repair enclosure. Not only does this increase the duration of the repair,
sinee injection ports can be predrilied on a flange repair enclosure, but
also the likelihood of damaging a stud.

Seal ant is injected in one-cubic-inch slugs using a hydraulically-operated

ram and cyclinder threaded into the injection valve. The ram can be either
hand-pumped or activated by air for 1arger sealant volumes. Final
reinjections are always hand-pumped with hydraulic tool injection pressures
ranging from 3000-4000 psi (20.7-27.6 mPa). The actual injection sequence
used during the repair of the 4 in (100mm) flange is given below (refer to
Figure 5-4).

5-4
Injection Number Injection Port Vents

1 4 1,2,6,7
2 5 1,2,3,6
3 3 1,6,7,8
4 6 1,2,7,8
Port 6 drilied out
5 6 1,2,7,8
6 7 1,2,8
Port 7 drilled out
7 7 1,2
8 6 1,2
Port 3 drilied out
9 3 1,2
Port 8 drilled out
10 8 1,2
11 2 1
Flange face peened to reduce leakage
12 1 No vent
Port 1 drilied out

The original leak in this case was between ports 8 and 1. Note that ports
are drilied out only for injection purposes, never to reestablish a vent.
The final injections were made with no vent to atmosphere except through the
path of leakage. Single cavity enclosures were also used on all 3/4 in
(19mm) pipe specimens (see Figure 5-7). These enclosures contained two
injection ports, one located directly over the leak, and the other 180°

apart. Final injections were made with no vent, and in one instance a final
injection was made with the leak already sealed. Observation of Vendor C
repairs led one to believe that their procedures do not call for any
diffusion of the leak prior to the installation of the repair enclosure. On
the first attempt by Vendor C to seal a pipe specimen, stainless steel wire
was wrapped around the pipe in the area of the leak; this had no effect on
the magnitude of the leak. Main stream extrusion was significant, and upon

5-5
disassembly it was evident that the wire had been completely displaced
during the installation of the repair enclosure. After Vendor C personnel
were made aware of this extrusion, a hose clamp was used for the final pipe
test to diffuse the leak, requiring a portion of the repair enclosure cavity
to be ground away to accommodate it. Prior to this, two other pipe leaks
were sealed with no attempts made at diffusing the leaks.

Inasmuch as expulsion of seal ant from repair fixtures is common during

injection, the fact that Vendor C can account for the total volume of
sealant used is of little significance. In practice, the use of an
atmospheric vent permits the escape of seal ant, thus 1imiting the
effectiveness of predetermined volume control. If atmospheric vents are not
employed, then system pressure must be exceeded during injection, thereby
increasing the chance of seal ant extrusion into the process stream.

Vendor C valve repairs were accomplished without the use of enclosures.

Stuffing boxes and bonnet gaskets were drilied, tapped, and injected
directly.

5-6
Figure 5-1. Vendor A Enclosure on 4 in Flange

5-7
Figure 5-2. Comparison of 3/4 in Mange Repair Fixtures
Figure 5-3. Vendor B 2 in Valve Repair - Note: Packing Repair
Successful; Existing Leak is in Bonnet Flange

5-9
5-10
5-11

Figure 5-5. Comparison Of 3/4 in Line Enclosures

 Section 6

TEST RESULTS AND DISCUSSION

EFFECTIVENESS OF REPAIRS

Table 1 compares the relative effectiveness of each vendor's leak repairs

after the test specimens had been exposed to line conditions for 48 hours.
Leak rates were determined by hydrostatic tests at ambient temperature and
2200 psi (15.2 mPa) pressure (when possible). Premature failures or
excessive leakage which prevented the full pressure from being obtained are
so noted. In addition, leak rates for the 6 in (150 mm) valve after 2 weeks
on-line are included.

Tables 2, 3 and 4 show test results from the long-term thermal cycling tests
performed on specimens sealed by Vendor A, Vendor B and Vendor C,
respectively. It is obvious that for the sealing conditions tested no
vendor can claim consistent success. Assuming an acceptance criterion for a
successful repair of zero leakage initially, then both Vendor A and Vendor B
can claim a success rate of about 50%, with 6 and 5 of 12 specimens
perfectly sealed, respectively; Vendor C's repairs showed zero leakage in
only two specimens. If an arbitrarily selected small leakage of one gallon
per day (approximately 160 cc/hr) is acceptable, then all vendors exhibited
an improved success rate : 7 out of 12 for Vendor A , 6 for Vendor B, and 5
for Vendor C.

The longest term over which the repairs were tested on line was 14 weeks;
the specimens tested were two 3/4 in (19mm) pipes and a 3/4 in (19mm)
flange. With the exception of the Vendor C specimen which suffered the bulk
extrusion, all the pipe repairs were successful both initially and over the
14 weeks. Al1 the 3/4 in (19mm) flange repairs held for the initial 48

6-1
hours, but the Vendor A specimen ultimately developed significant leakage
(>10,000 cc/hr) by the end of the test period. In addition, the 4 in
(100mm) flanges were to have remained on 1ine for 14 weeks; none lasted more
than 8 weeks before serious leakage occurred.

Repairs to valves were more random in their effectiveness. Bonnet leaks in

particular proved difficult to seal effectively; Vendor A, however, had good
success in sealing packing leaks. Vendor B failed on this portion of the
test, with all three valves leaking seriously (>10,000 cc/hr) after two
weeks on line. For Vendor A, the 3/4 in (19mm) valve specimen was their
only success; only Vendor C's repair of the 2 in (50mm) valve failed to seal
effectively. Observations of the valve repairs during the long-term tests
indicate that valve operation (opening and closing) had little effect on the
repairs. Thermal cycling, however, seemed to adversely affect the repaired
valves, as it did the f1ange and pipe specimens as well.

Table 5 shows the performance of sealed leaks during progressive stages of

the oven aging test (described in detail in Appendix B). These test
specimens were sealed while on-line, but aging was performed off-line in
ovens, in air. While not truly representative of long-term, on-line sealant
performance, the data nonetheless reflect comparative seal ant behavior under
controlled conditions of accelerated aging.

RELATIVE ADVANTAGES AND DISADVANTAGES

Vendor A

The quality of repair, longevity and insensitivity to thermal and mechanical

eye 1ing are relatively good. Adequate precautions are taken to reduce the
possibility of seal ant extrusion; whenever possible the leak is mechanically
reduced dramatically or stopped completely prior to injection. However, if
mechanical diffusion is not possible, such as with the bonnet leak in the 2
in (50mm) valve, the risk of extrusion is greatly increased. Figure 6-1
clearly shows the evidence of seal ant extrusion into the valve body.

6-2
Permanent repair of flanges and valves sealed by Vendor A could be
prohibitive, due to both sealing techniques and the sealant itself. The
liberal peering of flanges prior to sealing caused damage which would
require remachining of all flange faces sealed by Vendor A , if reuse was
contemplated. Typical damage is illustrated in Figures 6-2 and 6-3.
Disassembly of flanges and valves was extremely diffcult, due to the
concrete-like nature of the cured sealant. Although it does not adhere well
to metallic surfaces, its cohesive and incompressible nature made necessary
the cutting of valve studs and the use of a hydraulic press to remove studs
from flanges.

Vendor B

While the quality of the initial repair performed by Vendor B is comparable

to that of Vendor A, flange and especially valve repairs show a definite
deterioration under thermal cycling conditions. Pipe repairs utilizing the
Vendor B line enclosure (refer to Figure 5-5) did not display this
sensitivity, however.

Vendor B's approach to preclude sealant extrusion is built into their

injection procedure, by monitoring sealant injection pressure at the
injection valve, and not allowing the pressure to exceed system pressure.
However, even this method does not preclude extrusion; Figure 6-4 shows a
smal1 amount (< lOcc) of seal ant found in the downstream strainer following
the Vendor B repair of a 3/4 in (19mm) f1ange at 200°F (93°C).

Permanent repair of parts sealed by Vendor B should not pose any exceptional
problems. Flanges were peened at their interface with the repair clamp, and
not at the mating surfaces, thereby eliminating a source of flange damage.
The cured seal ant is nonadhesive and remained relatively soft, and did not
hinder the disassembly of any test specimens.

6-3
Vendor C

The initial effectiveness of the Vendor C repairs was marginal. The 3/4 in
(19mm) pipe specimens proved to be especially difficult to repair; Figure
6-5 shows a typical leak after the Vendor C repair failed. Of repairs that
held up for the initial 48 hours, however, most continued for the duration
for the long-term tests with little or no increase in leakage.

Vendor C's demonstration of leak repairs also resulted in the most severe
case of extrusion witnessed during the test program. This occurred during
the repair of a 3/4 in (19mm) pipe specimen at 600 F (316 C). The
initial repair failed after approximately one hour on-line. The repair
fixture was removed, cleaned, reinstalled and finally reinjected. After 48
hours on-line, when the test specimen was removed for hydrostatic testing,
two separate slugs of seal ant were discovered in the pipe downstream of the
repair. One slug was caught in a 90° pipe bend about 4 in (10 cm) away

from the test specimen, and the other slug was lodged in the test pipe
itself. Both slugs are shown in Figure 6-5. The test specimen was
sectioned for examination, and is shown in Figure 6-7. Note in the
photograph the displaced wire wraps in the vicinity of the leak, as referred
to in Section 5. From examination of the slugs and approximating their
3
volume as 0.5in (8 cc), it appears that the extrusions were produced
separately, one for each attempted repair.

Disassembly of sealed parts and removal of cured sealant for permanent

repair is time consuming but otherwise not a problem. Flange repairs,
though, may be affected by Vendor C's procedure which calls for dri11ing
into the leaking flange itself in order to install the injection valves.
These holes presumably would require permanent filling (i.e., weld repair),
and any studs damaged during drilling would have to be replaced.
 Table 1

INITIAL LEAKAGE AT 48 HOURS

cc/hr (gal/hr)

LINE CONDITIONS
TEST SPECIMEN I.D.# DURING SEALING VENDOR A VENDOR B VENDOR C
3/4" Pipe Specimens:
with 1/64" dia. hole 1 540°F, 2200 psi 0.0 (0.0) 0.0 (0.0) 4,360 (1.15)
200°F, 2200 psi ^51,000 (13.47)
with 1/32" dia. hole 2 0.0 (0.0) 0.0 (0.0)
with 1/16" dia. hole 3 600°F, 2200 psi 0.0 (o.o) 0.0 (0.0) (2)

with 1/16" dia. hole 4 200°F, 2200 psi 0.0 (0.0) 0.0 (0.0) 10 (<0.01)
6" Valve 5 600°F, 2200 psi 2,720b (0.72) 8,000a (2.11) 10afb (<0.01)
cr> 3/4" Valve 6 600°F, 2200 psi 0.0 (0.0) 240a (0.06) (3)
^ 2" Valve 7 200°F, 2200 psi 0.0 (0.0) 740b (0.19) 40b (0.01)
3/4" Flange 8 200°F, 2200 psi 80 (0.02) 0.0 (0.0) 0.0 (0.0)
3/4" Flange 9 600°F, 2200 psi 200 (0.05) 1,120 (0.29) 6,000 (1.58)
3/4" Flange 10 600°F, 2200 psi 3,440 (0.91) 120 (0.03) 13,000 (3.43)
3/4" Flange 11 600°F, 2200 psi 13,240 (3.49) 1,600 (0.42) 19,000 (5.02)
4" Flange 12 600°F, 2200 psi 72,314 (10.10) 9,000 (2.37) 5,000 (1.32)
6" Valve Leakage after 2 wks on-1ine 9,720b (2.56) 13,200a(3.48) 100a,b (0.02)
(1) Removed from line conditions after 24 hrs; 1500 psi was maximum hydrostatic pressure attainable
(2) Could not be hydrostatic tested without disturbing bulk extrusion (see text)
(3) Negligible: leakage from packing, about 1 drop/min.
(a) Packing leak

(b) Bonnet flange leak

 Table 2

RESULTS OF LONG-TERM THERMAL CYCLING TESTS

VENDOR A

co 00 CO ^
2: O CO 2: Q CO 2: 0 CO
0 UJ >- O UJ >- O UJ >-
LEAKAGE HH s: LU LEAKAGE t“-i UJ LEAKAGE (—1 UJ LEAKAGE
{
— C£ ^ ^
cc: 1—QC ^
LINE CONDITIONS AFTER C O AFTER <c 0 AFTER < O AFTER
01 U- LU u.
cc: uj QC U- UJ
DURING 48 HR LONG-TERM LU LU 2 WEEKS UJ OC UJ 8 WEEKS UJ DC UJ 14 WEEKS
Q- UJ CO A- uj co CL UJ CO
SPECIMEN # SEALING cc/hr LINE CONDITIONS 0 a. ^ cc/hr So- — cc/hr O CL'— cc/hr

1 540°F, 2200 psi 0 540°F, steam 1 - 2 0 - -

3 600°F, 2200 psi 0 600°F, 2200 psi 1 0 2 0 1 0

4 200°F, 2200 psi 0 600°F, 2200 psi 1 0 2 0 1 0

5 600°F, 2200 psi 2720b 600°F, 2200 psi 6 9720b - - - -

6-6

6 600°F, 2200 psi 0 600°F, 2200 psi 5 0 6 3a - -

7 200°F, 2200 psi 0 600°F, 2200 psi 5 56 b 6 1100b - -

8 200°F, 2200 psi 80 600°F, 2200 psi 1 6900 3 5200 4 10,580

12 600°F, 2200 psi 72,314 600°F, 2200 psi 1 X - - - -

X = test terminated due to excessive leakage KEY TO OPERATIONS PERFORMED

1- 2 temperature and pressure (T/P) cycles to ambient
(a) Packing leak
2- 4 T/P cycles to ambient
(b) Bonnet flange leak
3- 6 T/P cycles to ambient
4-8 T/P cycles to ambient
5-4 valve cycles hot, 2 valve cycles @ ambient
6-4 valve cycles hot, 2 valve cycles 0 ambient,
2 T/P cycles to ambient
 Table 3

RESULTS OF LONG-TERM THERMAL CYCLING TESTS

VENDOR B

OPERATIONS

OPERATIONS

OPERATIONS
(SEE KEYS)

(SEE KEYS)

(SEE KEYS)
PERFORMED

PERFORMED

PERFORMED
LEAKAGE LEAKAGE LEAKAGE LEAKAGE
LINE CONDITIONS AFTER AFTER AFTER AFTER
DURING 48 HR LONG-TERM 2 WEEKS 8 WEEKS 14 WEEKS
SPECIMEN # SEALING cc/hr LINE CONDITIONS cc/hr cc/hr cc/hr

1 540°F, 2200 psi 0 540°F, steam 1 - 2 0 - -

3 600°F, 2200 psi 0 600°F, 2200 psi 1 0 2 0 1 0

4 200°F, 2200 psi 0 600°F, 2200 psi 1 0 2 0 1 0

5 600°F, 2200 psi 8000a 600°F, 2200 psi 6 13,200a

6-7

6 600°F, 2200 psi 240a 600°F, 2200 psi 5 18,925a 6 34,069a

7 200°F, 2200 psi 740b 600°F, 2200 psi 5 82,270b

8 200°F, 2200 psi 0 600°F, 2200 psi 1 0 3 6 4 0

12 600°F, 2200 psi 9000 600°F, 2200 psi 1 84,000 2 X - -

X = test terminated due to excessive leakage KEY TO OPERATIONS PERFORMED
1- 2 temperature and pressure (T/P) cycles to ambient
(a) Packing leak
2- 4 T/P cycles to ambient
(b) Bonnet flange leak 3- 6 T/P cycles to ambient
4 - R T/P cycles to ambient
5- 4 valve cycles hot, 2 valve cycles @ ambient
6- 4 valve cycles hot, 2 valve cycles @ ambient,
2 T/P cycles to ambient
 Table 4

RESULTS OF LONG-TERM THERMAL CYCLING TESTS

VENDOR C

OPERATIONS
OPERATIONS

(SEE KEYS)
OPERATIONS

(SEE KEYS)
(SEE KEYS)

PERFORMED
PERFORMED
PERFORMED
LEAKAGE LEAKAGE LEAKAGE LEAKAGE
LINE CONDITIONS AFTER AFTER AFTER AFTER
DURING 48 HR LONG-TERM 2 WEEKS 8 WEEKS 14 WEEKS
SPECIMEN # SEALING cc/hr LINE CONDITIONS cc/hr cc/hr cc/hr

1 540°F, 2200 psi 4360 540°F, steam 1 X - - -

600°F, 2200 psi -k 600°F, 2200 psi 1

3

4 200°F, 2200 psi 10 600°F, 2200 psi 1 <2 2 24 1 2,5

6-8

5 600°F, 2200 psi 10a,b 600°F, 2200 psi 6 100a,b - - - -

6 600°F, 2200 psi <10a. 600°F, 2200 psi 5 36a 6 60a - -

200°F, 2200 psi 40b 600°F, 2200 psi C 10,920b X - - -

7

8 200°F5 2200 psi 0 600°F, 2200 psi 1 <2 3 0 4 0

600°F, 2200 psi 600°F5 2200 psi 1 2 35,500 X -

12 5000 i 18,000

X = Test terminated due to excessive leakage KEY TO OPERATIONS PERFORMED

1 - 2 temperature and pressure (T/P) cycles to ambient
* Bulk extrusion; see text
2 - 4 T/P cycles to ambient
(a) Packing leak 3- 6 T/P cycles to ambient

(b) Bonnet flange leak 4- 8 T/P cycles to ambient

5- 4 valve cycles hot, 2 valve cycles @ ambient
6- 4 valve cycles hot, 2 valve cycles 8 ambient,
2 T/P cycles to ambient
Figure 6-1. Extrusion Of Seal ant Into Body Of 2 in Valve Sealed By Vendor A

6-9
 Figure 6-2. Peem'ng Damage on 4 in Flange Sealed
By Vendor A

Briaa—
Figure 6-3. Close-Up Of Peening Damage On 4 in Flange Sealed Vendor A

6-10
Figure 6-4. Sealant Extrusion Found In Pipe And Strainer Downstream Of 3/4 in
Flange Sealed By Vendor B

6-11
Figure 6-5. Leak In Vendor C Line Enclosure Prior To Termination Of Test

6-12
Figure 6-6. Vendor C Line Enclosure With Extruded Sealant Slugs

6-13
6-14

Figure 6-7. Section Through 3/4 in Pipe Sealed By Vendor C, Showing Bulk Extrusion
 Section 7

MECHANICAL SEALING TEST

During the course of the leak sealing program, both Vendor A and Vendor B
suggested that the type of repair fixture used to repair pipe leaks could be
successfully applied to leaking valves and flanges. The general design
features of such a fixture include a void chamber to accommodate the flange
(or valve), and a perimeter seal on the repair fixture itself. This repair

concept presumably reduces significantly the risk of extruding sealant into

the process stream.

The three vendors participating in the leak sealing tests, plus a fourth
vendor, were invited to demonstrate their techniques for "mechanical11 or
"semi-mechanical" repair of on-line leaks. Test requirements were that one
3/4 in (19mm), 1500-pound flange would be sealed at line conditions of
620°F and 2200 psi (327°C, 15.2 mPa), and that the injection of sealant

between the two flange faces would not be necessary. Given these
conditions. Vendor B and Vendor A agreed to participate, while Vendor C
declined, citing limitations of this type of repair under the ASME Boiler
and Pressure Vessel Code. A fourth vendor. Vendor D, also agreed to
participate. The individual repairs are discussed separately below.

VENDOR A

The repair fixture used by Vendor A is shown installed in Figure 7-1. Prior
to assembling the fixture around the flange, Vendor A followed their
previously demonstrated procedure of first diffusing the leak by peem'ng the
flange. This was followed by wrapping the bolt circle with stainless steel

7-1
wire until it approximately equal led the f1ange O.D. The fixture was then
assembled about the flange, and pumped with sealant. The duration of this
initial repair was about 6 hours. After 14 hours, the fixture was repumped
with sealant for 2 hours.

After 48 hours at line conditions, the fixture was taken off-line, cooled
and tested for leaks. Zero leakage was measured at 2200 psi (15.2 mPa).

The disassembled fixture is shown in Figure 7-2. As noted previously with

Vendor A-sealed test specimens, disassembly was difficult due to the
concrete-1 ike condition of the cured seal ant. The entire void around the
flanges was filled with sealant, and further disassembly to expose the
flange (which was not attempted) could easily have resulted in damage to the
flange, the pipe, or both.

VENDOR B

The Vendor B fixture, as shown assembled in Figure 7-3, was the smallest and
the lightest (50 lb (22.7 kg)) fixture demonstrated, and was the only one
with a single perimeter seal. Duration of initial repair was 3 hours, and
repumping for 1 hour was performed 10 hours after the initial repair. After
approximately 36 hours on-1ine, a significant leak occurred near a stud, and
continued until the fixture was taken off-1ine. (Under normal circum­
stances, al1 vendors are available for a repump within 48 hours of the
initial repair. Hoi iday schedules of both Vendor B and C-E eliminated the
opportunity for a 1 ate repump.) Leak testing at ambient temperature showed
the leak rate to be 12,800 cc/hour at 2200 psi (15.2 mPa).

When the Vendor B fixture was disassembled, the reason for the substantial
leak was discovered. Again, due to hoiiday schedules and the desire to
complete testing prior to the end of the calendar year, the Vendor B
enclosure was designed from a drawing of the flange and pipe, rather than

7-2
from actual field measurements. The pipe actually tested had an O.D. 5 to
10 mils (0.13-0.25 mm) 1arger than that specified in the drawing sent to
Vendor B. This mismatch of pipe and fixture prevented f1 at contact between
the upper and lower flange faces of the repair fixture, especially near the
pipe/fixture intersection, thereby shortening the path from void to studs.
A photograph of the disassembled fixture is shown in Figure 7-4, and the
flange is shown removed from the fixture in Figure 7-5.

VENDOR D

A communication problem undoubtedly exacerbated by the accelerated test

schedule resulted in the delivery to C-E of a grossly oversized repair
fixture, shown in Figure 7-6, weighing 652 pounds (296 kg). Vendor D's
request for additional time to manufacture an appropriately-sized fixture
had to be denied due to schedule pressures. Since a new fixture would
differ only in exterior dimensions (and weight), it was agreed that the
fixture on hand would be tested to evaluate Vendor D's concept.

The Vendor D mechanical sealing fixture was the only one tested which
incorporated separate, inner and outer (removable asbestos) seals in
addition to an intermediate groove for seal ant containment. This design
feature serves as a reasonably effective means of containing seal ant away
from the leaking flange, helping to insure that no seal ant is extruded into
the process stream. The photographs of the disassembled fixture in Figure
7-7 clearly show the absence of seal ant in the void chamber.

The f1ange repair with the Vendor D fixture failed for the same reason as
did the Vendor B unit, which was an over-sized pipe with respect to the pipe
clearance manufactured in the fixture. As with the Vendor B fixture, true
pipe dimensions 1arger than those stated on the drawing provided to Vendor D
were responsible. The net result was a cold leak rate in excess of 15,000

7-3
cc/hour at 1800 psi (10.2 mPa). Vendor D was willing to provide a new
fixture of the correct size for this application, accommodating the true
pipe diameter, but additional test time was not available.

It should be noted that Vendor D is primarily concerned with selling leak

repair fixtures, and normally does not perform the actual sealing work.
Installation instructions and sealant are typically provided with the
fixtures, with on-site technical supervision available if required. In the
case of the mechanical sealing test, all sealing work was performed by
Vendor D employees.

SUMMARY AND CONCLUSIONS

Vendor A demonstrated the only successful flange repair utilizing the

mechanical repair fixture concept. However, it would not be prudent to draw
negative conclusions regarding the performance of the Vendor B and Vendor D
mechanical sealing fixtures. Both vendors' fixtures avoid packing the
flange with sealant, which has the important advantage of almost
guaranteeing no sealant extrusion into the process stream. A secondary
advantage is the relative ease of disassembly for permanent repair, if the
flange is not encased in cured sealant. However, the effectiveness of the
repair is contingent on the perimeter seal, and no definitive test results
were obtained for either fixture which relies exclusively on the perimeter
seal. By inspection, it appears that the Vendor D fixture, with its
additional inner and outer asbestos seals, offers an advantage over the
Vendor B fixture which relies on a single perimeter seal. The Vendor A
fixture and technique, while effective, exhibited a disadvantage in that
subsequent removal of the repair fixture is difficult due to the condition
of the cured sealant.

7-4
Figure 7-1. Vendor A Mechanical Sealing Fixture, Installed

7-5
Figure 7-2. Vendor A Mechanical Sealing Fixture, Disassembled After Test

7-6
Figure 7-3. Vendor B Mechanical Sealing Fixture

7-7
Figure 7-4. Vendor B Mechanical Sealing Fixture, Disassembled After Test
7-9
Figure 7-6. Vendor D Mechanical Repair Fixture, Installed
7-11
 Section 8

CONCLUSIONS AND RECOMMENDATIONS

It is assumed that each contractor participating in the leak repairs test

program provided their most experienced technicians, best available fixtures
and nuclear grade sealants. Therefore the test results obtained should be
representative of field performance. Based on the test data and
observations during the course of the program, it is apparent that none of
the participants offers a clearly superior technique or system for sealing
leaks on line. Despite the variation in sealing methods, all experienced
failures as well as perfect repairs and each vendor had at least one
incident of sealant extrusion into the process stream.

It was not the objective of the program to determine which, if any, of the
vendors provided the most effective and consistent leak repairs. Rather,
the demonstrations by the participating vendors and subsequent long-term
tests offered substantial insight into the state of the art regarding such
repairs. As such, the program allows the following conclusions and specific
recommendations to be made.

CONCLUSIONS

1. Sealant behavior during injection is unpredictable and difficult to

quantify. Off-gassing with subsequent spontaneous expulsion and
extrusion of sealant occurred several times during the performance of
these tests. The curing requirements vary among sealants, but vendors
demonstrated reasonable knowledge of their sealants' performance and
capabilities.

8-1
2. Sealant extrusion into the process stream was infrequent, but
observations made during the test program suggest that such extrusions
can not be positively prevented. Mechanical sealing fixtures (as
described in Section 7) may significantly reduce the risk of such
extrusion.

3. Diffusion of leaks by mechanical means such as peem'ng or wrapping with

wire or clamps does not appear to affect the ultimate effectiveness of
the repair, although it may inhibit sealant extrusion into the stream
and simplify the sealing process.

4. Limiting sealant injection pressures to system (line) pressure appears

to be an effective technique for inhibiting sealant extrusion. Viscous
and semi-solid sealants require hydraulic injection pressures
significantly in excess of system pressure, which provides little
control sensitivity for preventing extrusion. Further, injection
pressures for these sealants must be determined indirectly from the
injection tools, and not from the sealant injection line itself. Less
viscous (fluid) sealants may be injected at or near system pressure, and
also allow for monitoring of pressure directly.

5. The frequent explusion of sealant during the repair process suggests

that knowledge of total sealant volume injected (pre-determined volume
control) is a useful guideline, but is not a guarantee of prevention of
sealant extrusion.

6. All sealants and repairs showed deterioration under thermal cycling

conditions. It may be inferred with reasonable certainty that
successful repairs which failed with repeated thermal cycles would have
remained effective under steady-state conditions.

7. Valve repairs were generally less successful than repairs of pipes and
pipe flanges. This may be attributable to the fact that most of the
valves were repaired without the use of enclosures, but rather by
injecting sealant or pumpable packing material into the stuffing boxes
directly. The effectiveness of these repairs is thus more closely
related to the performance of the sealant (or pumpable packing material)
itself.

8-2
RECOMMENDATIONS

1. Customers employing leak repair vendors should require certified,

independent chemical analyses of each batch of the sealants used, to
determine the presence of reactive elements potentially harmful to the
system being sealed.

2. A repair with significant leakage within 48 hours of the completion of

the repair should be considered a failure, and a new repair initiated.

3. The added mass of repair fixtures may compromise the seismic

qualification of any piping systems employing such repairs (nuclear
plants). Seismic requirements should be considered prior to repair.

4. The design basis for the fixtures employed should be considered by the
user as these sometimes rather massive structures do become a new
pressure boundary once the repair is effected. Further, if the repair
fixture captures injected sealant between flange faces, the additional
loading on the flange studs due to injection pressure (above system
pressure), possible increase in effective sealing diameter, and
mechanical sealing operations must be considered.

5. Diffusion of flange leaks by peening flange faces should be limited to

minimize the complexity of subsequent permanent repair.

6. Al1 leak repairs utilizing an injected seal ant should be considered

temporary, and permanent repairs should be made as soon as practical.

7. The use of an additive to speed up curing should be evaluated for

specific applications.

8-3
 Section 9

REFERENCES

1. "Evaluation of Methods For Sealing Leaks During Plant Operation,"

C-E Document No. 00000-ESE-269, issued October, 1979.

2. C-E Drawing No. SC-12278-500-003, Rev. 01, 4 inch Flange Test

Sample Layout.

3. C-E Drawing No. SC-12278-500-005, 3/4 inch Tube Test Samples.

4. C-E Drawing No. SC-12278-500-008, 3/4 inch Flange Test Semple.

5. C-E Drawing No. SD-12278-500-009, Test Specimen, Sealant Extrusion

Assembly.

6. C-E Drawing No. SD-12278-500-010, 3/4 inch Forged Univalve Assembly.

7. C-E Drawing No. SD-12278-500-011, 2 inch Bolted Bonnet Globe Valve.

8. C-E Drawing No. SE-12278-500-012, 6 inch Forged Bolted Bonnet Valve.

9-11. Video Tape Recordings of Vendor A Leak Repairs, Tapes A-l, A-2, and
A-3.

12. Video Tape Recording of Vendor B Leak Repairs, Tape B-l.

13. Video Tape Recording of Vendor C Leak Repairs, Tape C-l.

9-1
 APPENDIX A

VENDOR SURVEY

A cross-section of industrial customers of leak repair vendors was surveyed

to gain perspective on actual use of such repairs in the field. The survey
was conducted by telephone, and the respondents were mainly plant
maintenance engineers directly responsible for the selection of leak repair
services. Industries represented in this survey include chemical (2),
petroleum (5), steel (1) and electric power generation (6). Three of the
six electric utilities surveyed were nuclear generating stations.

Respondents were questioned about the following: which leak repair vendor
was selected; the frequency of use; the types of components sealed; line
conditions at the leak; and the general performance of the repair -
effectiveness, presence of stringers and/or effect on system chemistry,
difficulties with permanent repair, etc.

The components which most frequently required leak sealing were valves, from
1/2 in to 30 in (1.2 to 76 cm). Flanges were the next most common source of
leakage, with only a few users reporting pipe leaks. Line conditions for
the repairs ranged in temperature from 250°F to 1250°F (120°C to
675°C) and in pressure from 150 to 2600 psi (1 to 18 mPa). Both thermal

cycling to ambient temperature and pressure, as well as mechancial eye 1ing

of valves (opening and closing), were as common as steady-state conditions.

Respondents on the whole felt that vendor repairs were generally effective,
and they recognized the temporary nature (6-12 months) of the repairs.
Typically, 20% of all repairs required reinjection well after the initial

A-l
repair was made. High temperatures, high pressures, and especially thermal
cycling were cited as the most common causes of repair failure which
necessitated reinjection. However, some users reported valve repairs which
had lasted 2-1/2 years before failure.

No users reported any serious doubts about the chemical composition of the
sealants used, generally trusting the vendor's judgement as to the
suitability of the sealant used for the specific application. No chemical
analyses of sealant materials (to determine system compatibility) was
demanded of the leak repair vendors, or performed by the customers. Effects
of repairs on process (plant) chemistry were either unnoticed pr
nonexistent. However, about 20% of the respondents reported evidence of
stringers in their systems as a result of leak repairs.

In contrast to observations made during the leak sealing test program, the
removal of cured sealant from parts sealed by all three vendors in field
applications presented no special problems. In contrast, flanges and valves
sealed by Vendor A during the test program had proven to be almost
impossible to disassemble without damage.

It should be noted that only firms which had employed leak repair vendors
were included in the survey; some firms undoubtedly decline to inject
sealants to repair their leaks. With this in mind, it is perhaps not
surprising that all respondents felt the concept of on-line leak sealing was
cost effective, and that results were at least satisfactory. Vendor C
enjoys the largest share of the leak repair market among the users surveyed,
with Vendor A and Vendor B sharing the balance. Only two respondents had
used more than one vendor, so no conclusions can be drawn about customer
preference.

A-2
 APPENDIX B

SEALANT MATERIAL ANALYSES

Section B-1

IMPURITY CHEMISTRY OF THE SEALANTS

INTRODUCTION

The objective of the material analyses was to determine if elements were

present in the sealants used during the test program which could be
detrimental to the metallurgy or chemistry of reactor components or fuel, or
which would otherwise interfere with proper reactor operation. The sealants
were analyzed following each of these activities: thermal curing, gamma
radiation in primary coolant at ambient temperature, gamma radiation in
primary coolant at operating temperatures, and thermal aging.

PROCEDURES

The sequence of sealant tests was performed per the basic diagram in Figure
B-1. Individual test procedures are detailed in the following sections.

Curing of Seal ant

Samples of each vendor's sealant were cured for 20 hours at 475°F

(246°C) in a hot air oven. The samples were contained in aluminum

dishes, and were weighed before and after curing to determine weight
loss.

B-1
Irradiation of Sealant

Gamma irradiation of the sealants to 1x10^ rads was performed at

Neutron Products in Gaithersberg, Maryland. The sealant samples were

immersed in a solution of 6000 ppm boron as H^BO^ in deionized water
during irradiation. Samples irradiated at ambient temperature and
pressure were placed in open jars in 350cc of coolant. Radiation
exposure to a total of 1x10^ rads was carried at a dose rate of 1.17

Mrad/hr for 8.55 hours. Samples irradiated at primary conditions were

placed in stainless steel chambers with 150cc of coolant. The chambers
were heated by external band heaters. Temperatures were controlled from
thermocouples immersed in the coolant. During irradiation the chambers
were maintained at 650°F and 2200 psi (343°C5 15.2 mPa), except as

noted in the discussion of the individual sealants. The dose rate for
these samples was 0.667 Mrad/hr for 15 hours.

The primary coolant in which each of the sealant samples was irradiated,
retained and subsequently analyzed for Teachable elements.

Thermal Aging

Following irradiation, a portion of each sealant was thermally aged for

8 weeks at 400°F (215°C) in an air oven.

Elemental Analysis

The sealants and primary coolant samples were analyzed for elemental
composition at National Spectrographic Laboratories in Cleveland, Ohio.
Individual analysis techniqes are described as follows:

B-2
Total halogens in sealant: Samples were pyrolytically decomposed using
externally-heated furnace with an 0^ gas sweep into a liquid
caustic trap. Halogens were measured with specific ion electrodes per
ASTM Dll79-80 (B).

teachable halogens and total halogens in coolant: Samples were

extracted with deionized water for 4 hours at 203OF (95°C).

Halogens in the extract were measured by specific ion electrodes as

above.

teachable sulfur: Sulfur in the extract was determined

turbidimetrically per ASTM D516-80.

Phosphorus: Samples were analyzed for phosphorus colorimetrically

according to ASTM E350.

Mercury: Samples were analyzed for mercury by the Cold Vapor Generation
Method according to EPA 245.1.

Total lead, copper, zinc tin and cadmium: These metals were analyzed by
wet ashing around a gallium oxide carrier, followed by emission
spectroscopy.

Coolant Analysis

An analysis was run on the coolant used for the sealant environment
during irradiation. Analytical measurements were made on all coolant
samples, including the blank, after irradiation. Except as noted in

B-3
 Table B-3, no increases in the elemental impurity levels of the coolant
samples, as compared with the original solution, were measured. This
indicates that no measurable quantities of impurities (except as noted)
were extracted into the coolant during irradiation.

RESULTS AND DISCUSSION

Vendor A

Test results for this sealant are given in Table B-1.

This material was stable with respect to heat both during the initial cure
(20 hours at 475°F) and the 8 week, 400°F thermal aging following

irradiation. In terms of elemental impurities, the chlorine levels exceed

the limits for non-metallic materials established by Combustion Engineering,
Inc., for use in its nuclear plants (Table B-4). (NOTE: Impurity limits in
Table B-4 are given for reference purposes only.) No other impurities were
present above the levels of detectability.

Vendor B

Test results for this sealant are given in Table B-2.

The Vendor B sealant exhibited very poor stability during the thermal
(curing and aging) cycles. While curing, this material lost over 30% of its
initial weight, a brown viscous oil distilled out of the sample, and a great
deal of shrinkage was noted. During the post-irradiation thermal aging, an
additional weight loss of 22% was measured, resulting in a total weight loss
of over 50%.

B-4
Regarding impurities, the Vendor B sealant has very high levels of chlorine
and sulfur, and high levels of fluorine. Significant amounts of both
chlorine and fluorine also leached into the cool ant during irradiation at
primary conditions.

Vendor C

Test results for this sealant are given in Table B-3.

During the initial cure, the Vendor C sealant gained weight, indicating that
an oxidative curing reaction may have occurred. The weight change during
the 8-week thermal aging showed a moderate loss.

Elemental impurity levels were general ly low, although the chlorine levels
are above Combustion Engineering criteria (refer to Table B-4). No other
impurities were present above the 1imits of detectability.

B-5
 Table B-1

SEMjANT MATERIAL ANALYSIS: VENDOR A

I Weight change during cure (20 hours at 475°F) = -0.73%

II Elemental analysis - reported in ppm „, „(*)

Irradiated Irradiated
(1x10? rads) (1x10^ rads)
Limit
of Cured ^ Ambient Primary
Element Detectability Unaged Aged Unaged Aged Unaged Aged

Total F * 76.1
(2) k k
10 (2)

Total Cl 10 730 761 (2) 561 (2) 240

Total Br 10 * k
(2) k
(2) k

Total I 10 * k
(2) k
(2) k

Leachable F 10 * k k k k k

Leachable Cl 10 * 167 k
165 k k

Leachable S 100 * * k * k k

P 10 •k k k * k k

k k k k k k
Eg 0.5

k k k k k k
Pb 10

k k k k k
Cu 10 k

k k k k k
Zn 10 k

k k k k k
Sn 10 k

k k k k k k
Cd 50

(*) Below limit of detectability of test

(1) Aged 8 weeks at 400°F

(2) Equipment failure prevented measurement

III Weight loss after heat aging (8 weeks at 400°F) =1.0%

B-6
 Table B-2

SEALANT MATERIAL ANALYSIS: VENDOR B

I Weight change during cure (20 hours at 475°F) = -33.0%

Dark brown oil distilled out of sample during aging(*)

II Elemental analysis - reported in ppm Irradiated Irradiated

(1x10^ rads) (1x107 rads)
Limit
of Cured ^ Ambient Primary
Elemental Detectability Unaged Aged Unaged Aged1 Unaged Aged

10 175 2354
Total F (2) 250 (2) (2)
4 1130'
Total Cl 10 1960 1550 (2) 720 205

Total Br 10 (2) *
(2) * (2) *

Total I 10 (2) *
(2) k
(2) *

Leachable F 10 (2) (3) k

100 * (3)

Leachable Cl 10 710 (3) 745 1470 165 (3)

Leachable S 100 2380 (3) 2610 (3) 2300 (3)

P 10 * 100 30 200 60 180

0.5 ■k * k * * *
Hg

Pb 10 k * k * k k

k k
Cu 10 21 ■k
20 30

Zn 10 * k k * k k

Sn 10 * k k k k k

* k k k k k
Cd 50

(*) Below limit of detectability of test

(1) Aged 8 weeks at 400°F
(2) Equipment failure prevented measurement
(3) Insufficent sample
(4) Chlorine level in coolant is high indicating chlorine leached into coolant
during irradiation. The chlorine level in coolant blank is 32 ppm - after
primary irradiation. Cl-in coolant is 58 ppm. Fluorine level in coolant is
high indicating fluorine leached into tie coolant. The fluorine level in the
coolant blank is 6 ppm - after primary irradiation the fluorine level in
coolant is 32 ppm.
Ill Weight loss after heat aging (8 weeks at 400°F) = 22%

B-7
 Table B-3

SEALANT MATERIAL ANALYSIS: VENDOR C

I Weight change during cure (20 hours at 475^F) = +4.6%

II Elemental Analysis - reported in ppm(*)

Irradiated Irradiated
Limit (1X107 rads) (1x10^ rads)
of Cured Ambient „ Ambient
1"
Element Detectability Unaged Aged Unaged Aged Unaged Aged

Total F 10 (2) ■k * k
(2) *

Total Cl 10 (2) 251 (2) 210 (2) 376

Total Br 10 (2) * (2) * (2) *

Total I 10 (2) * (2) *

(2) k

Leachable F 10 * * * * k k

Leachable Cl 10 * 179 * 183 k

134

Leachable S 100 * * * * k *

P 10 * * * * k *

Hg 0.5 ■k * * k k k

Pb 10 * * * k k k

Cu 10 94 40 92 20 84 38

Zn 10 ■k * * k * k

Sn 10 19 * 15 k
14 k

Cd 50 * * ■k k * k

(*)Below limit of detectability of test

(1)Aged 8 weeks at 400°F

(2)Equipment failure prevented measurement

III Weight loss after heat aging (8 weeks at 400°F) =6.0%

B-8
 Table B-4

LIMITS OF ELEMENTAL IMPURITIES IN NON-METALLIC MATERIALS

ESTABLISHED FOR COMBUSTION ENGINEERING NUCLEAR STEAM SUPPLY SYSTEMS

Acceptance Criteria In ppm For

Non-Metallic Materials In or On:

Spent Fuel or
Impurity Reactor Intervals Steam Generator Refueling Pool

Total F 50 1000 50

Total Cl 100 100 100

Total Br 50 100 50

Total I 50 100 50

Leachable F 50-100* 10 10

Leachable Cl 50-100* 10 10

Total Sulfur 100 100 100

Leachable Sulfur 50 50 50

Total Mercury 0.5-50* 100 0.5

Total Lead 0.5-50* 50 0.5

Total Copper 200 100 50

Total Zinc 200 100 50

Total Tin 100 100 100

Total Cadmium 50 100 50

Total Phosphorus 1000 100 1000

*Depending on application

B-9
 CURE SEALANT
SPECIMENS„

IRRADIATE AT MBIENT IRRADIATE AT PRIMARY

oi-a

UNIRRADIATED CONDITION CONDITIONS GSOOF,

8.6 hr @ 1.2 Mrad/Hr 2200 psi 15 hr @ 0.7
SAMPLE Mrad/hr.

THERMAL AGE THERMAL AGE THERMAL AGE

8wks @400°F wks @ 400^F

ANALYSIS ANALYSIS ANALYSIS ANALYSIS ANALYSIS ANALYSIS

Figure B-1. Test Sequence For Sealant Samples

 Section B-2

THERMAL AGING OF SEALED FLANGES

INTRODUCTION

To gain additional insight regarding the nature and performance of the

sealants used during the leak repair evaluation program, long-term (8 weeks)
oven aging tests were performed on 3/4 in (19mm) flanges sealed on-line by
each vendor. Leak checks were performed periodically during the 8 weeks,
and samples of oven-aged sealant were tested for hardness at the same
intervals.

PROCEDURES

All flanges were sealed on-line with internal pressure and temperature of
the borated water test medium at 620°F, 2200 psi (327°C, 15.2 mPa).

After 48 hours on-line, the test specimens were cooled, depressurized, and
hydrostatically tested for leaks at room temperature. Flanges were aged in
groups of three at 300°F, 45Q°F, and 750°F (149°C, 232°C, and
398°C) for 8 weeks total. After the first two weeks, and at subsequent

two-week intervals, the flanges were again checked for leaks, and hardness
measurements were taken on samples of the sealant.

RESULTS AND DISCUSSION

Leakage measurements for the aged flanges are shown in Table B-5, and
hardness readings for the aged sealant specimens are given in Table B-6.

In general, the sealing performance of the repaired flanges was poor

throughout the period of oven aging. With few exceptions, leak rates
increased substantially with additional aging. It should be noted, however,
that the flanges were aged off-line, in air; this does not represent
"normal" conditions.

B-11
Sealant hardness measurements varied during the aging period. Only Vendor A
sealants uniformly increased in hardness at al1 temperatures. Seal ants from
Vendor B and C reacted similarly, increasing in hardness with aging at
300°F, but both curiously decreasing in hardness with aging at 450°F.

A uniform trend relating sealant hardness (or change in hardness) with the
rate of f1ange leakage was not clearly defined. It had been hypothesized
that as the sealant materials hardened with thermal aging, any elastomeric
properties would diminish, and leakage would increase. As stated above,
flange leak rates showed a general increase during the aging period, but
this trend does not appear to be associated with the change in seal ant
hardnesses.

B-12
 ft

Table B-5

SEALING PERFORMANCE OF 3/4-INCH FLANGES DURING OVEN AGING

Leak Rates After Given Duration at Temperature

cc/hr @ 2200 psi (except as noted)

Initial Leakage
Vendor Aging After 48 hrs.
& Specimen No. Temperature On Line 2 weeks 4 weeks 6 weeks 8 weeks

Vendor A 10 300° F 3440 9600 9100 8100 2700

11 450°F 13,240 14,400 6800 3000 9360

@ 2000 psi

9 750°F 200 40 0 0 0
B-1 3

Vendor B 9 300°F 1120 288 32 460 330

@ 2000 psi

10 450°F 120 17,160 7200 7200 18,000

@ 1000 psi @ 1000 psi 2000 psi

11 750°F 1600 17,040 38,760 >180,000 NA

@ 300 psi @ 300 psi @ 50 psi

Vendor C 9 300°F 6000 15,200 8250 7800 16,000

@ 600 psi @ 600 psi @ 2000 psi

10 450°F 13,000 24,720 24,000 29,700 36,000

@ 1400 psi @ 600 psi @ 1000 psi

11 750°F 19,000 43,320 >180,000 NA NA

@ 250 psi @ 50 psi
 Table B-6

SEALANT HARDNESS MEASUREMENTS DURING OVEN AGING

HARDNESS READINGS AFTER GIVEN DURATION AT TEMPERATURE

Vendor Agi ng
& Specimen No. Temperature 2 weeks 4 weeks 6 weeks 8 weeks Scale

Vendor A 10 300° F 52-55 47-63 73-76 68-72 Rockwel1 M

11 450°F 20-24 25-26 56-66 43-70 1/4" ball

9 750°F 39-50 54-58 53-70 63-65 100 kg load

B-14

Vendor B 9 300° F 75-80 85-90 85-90 85-90 Durometer,

10 450°F 75-80 65-70 55-60 35 Shore D

11 750°F 75-80 84-90 80-85 85-90

Vendor C 9 300° F 14.8-17.7 20.1 24.3 87.8 Microhardness,

0
10 450 F 12.6-13.1 9.1 8.9 3.8 Vickers DPH

11 750°F (1) - - -

NOTE:
(1) Sealant sample broke down into non-cohesive powder.