STEAM PLANT SURFACE CONDENSER LEAKAGE STUDY

EPRI IMP-481
(Research Project 624-1)

Final Report

Volume I

March 1977

Prepared by

BECHTEL CORPORATION
P. O. Box 3965
San Francisco, California 94119

PROJECT ENGINEER
Walter B. Lawrence

PRINCIPAL INVESTIGATORS
William D. Ellis
Frederik J. Hekking
Michel P. Lagache
Aswath M. Maddagiri
A. Charles Madsen

Prepared for

Electric Power Research Institute

3412 Hillview Avenue
Palo Alto, California 94304

EPRI Project Manager

William L. Lavallee

(M'
WSTRiEUliQNo,'
^'^OiViENT IS UNLIMITED
 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.
 FOREWORD

Tube failures in steam plant surface condensers are a significant reliability

problem which continues to confront the electric power industry. Part of the
impact of condenser leakage can be measured by the replacement power costs
experienced by large base load units as a result of outages to repair condenser
leaks. Another aspect of condenser leakage impact which is much more difficult
to measure is the potentially harmful effects of leakage on other major com­
ponents such as steam generators and turbines.

In an effort to provide the power industry with useful information to use in

assessing condenser leakage problems, the Electric Power Research Institute
contracted Bechtel Corporation to conduct this "Steam Plant Surface Condenser
Leakage Study."

There are numerous variables which can affect the leakage integrity of a con­
denser, too numerous for them all to be addressed with the limited amount of
data available to the writers of this report. These variables result from such
things as differences in plant operating philosophies, environmental differences
to which the condenser components are exposed, and differences in designs and
materials. Because of this, plant operators should view the conclusions reached
in this report as general considerations to be used as starting points for
reaching decisions on specific requirements for their condensers. The failure
rate probabilities presented in this report for various tubing materials serve
as a useful means for summarizing the tube failure experience of the units
participating in the field survey. Although care was taken in selecting the
sample of units surveyed as is explained in the report, no attempt has been made
to show that the sample is statistically representative of the entire population.
Such statistical treatment was not considered necessary since the failure rate
probabilities are not being portrayed as generic to the various tubing materials,
but only as a means for making relative comparisons among the materials surveyed.

Field surveys of this type are useful in collecting certain kinds of information,
but there are areas which are not readily assessable via this method. For example,
the report indicates that there were no demonstrated cases of severe damage to
other components in the steam cycle as a result of condenser leakage. Without
detailed failure analyses, the effect of condenser leakage on other components
can not be objectively verified, but there is considerable evidence to indicate
that such leakage does have detrimental effects specifically in the case of PWR
steam generators.

This report has demonstrated that operating power plant experience is extremely
useful in assessing equipment performance. However, this plant experience can
only be derived as a result of effective records keeping on the part of the
plants.

William L. Lavallee
Project Manager

iii
 ABSTRACT

This report presents the results of a study to determine the factors which
affect the deterioration and subsequent leakage of main surface condenser
tubes in electric power plants. The report addresses several areas related
to condenser tube leakage, including design, materials, chemistry, operation,
and maintenance. The data on which the report is based was obtained through
field trips to operating power stations, mail questionnaires, and a literature
search. The field survey concentrated on large fossil-fueled plants, nuclear
plants, and plants having condensers tubed with titanium. Much emphasis was
placed on presenting tubing failure information as a function of tubing
material, cooling-water conditions, failure mechanisms, condenser design, and
several other parameters. Information is also presented on the criteria for
retubing condensers, and recommendations for future work pertaining to con­
densers is discussed.

v
 ACKNOWLEDGMENTS

This study could not have been done without the interest and cooperation of the
organizations contacted by the study team. EPRI and Bechtel wish to acknowledge
the excellent cooperation afforded us throughout the industry. For their
generous contributions of time, records, and counsel, we are especially grateful
to the personnel of the utilities who operate the 30 stations that were surveyed
in the field.

The Bechtel study team was under the direction of Walter B. Lawrence, Project
Engineer, who was responsible for the conception and overall execution of the
study program. He prepared Sections 1, 2, 3, 6, 7, and the last two parts of
Section 5 for the final report.

William D. Ellis, Senior Engineer, as the full time participant of the study
team, was responsible for the day-to-day execution of the project and was the
lead engineer of the field survey teams. He also handled all computer
applications and data processing and contributed to all sections of the final
report as well as preparing the sections on titanium alloy tubes and records
keeping.

Michael F. Lagache and Aswath M. Maddagiri, Engineering Specialists, contributed

their engineering experience throughout the project. Both participated
heavily in the field survey work. Mike Lagache prepared Section 4.

A. Charles Madsen, Senior Metallurgist, gave expert guidance in his specialty

as well as full participation in the field surveys. He prepared most of
Section 5 of the final report.

Frederik J. Hekking, Senior Engineer, conducted all aspects of the literature

survey from the initial searches to the final printout of the bibliography
by the use of the computerized indexing system he devised.

vii
Ted N. Levine adapted the commercially available System 2K data management
software package to the requirements of the project, and instructed Bill Ellis
in its use.

Simone A. Linares, Project Administrator and Secretary, served with commendable

skill and resourcefulness.

The contract administrator for the first nine months was H. R. Mank who was
succeeded by William L. Cady. The two contract administrators conducted
all contractual matters including accounting and financial reporting.

Appreciation is also expressed to Mr. J. A. Prestele and Dr. Donald Anson for
their suggestions and comments on the preparation of this report.

viii
 CONTENTS

Section Page

1 INTRODUCTION 1-1
Statement of the Problem 1-1
Study Objective 1-2
Scope of Work 1-2

2 CONDENSER FUNCTION, DESIGN, AND OPERATION 2-1

Purpose of the Condenser 2-1
Condenser Configuration 2-2
Condenser Design Considerations 2-9
Maintenance of Tube Cleanliness 2-22

3 METHODOLOGY OF THE PRESENT STUDY 3-1

Introduction 3-1
Literature Survey 3-1
Matrix for Selection of Field Survey Stations 3-6
Mail Survey 3-7
Field Survey 3-10
Data Management System 3-19

4 CONDENSER OPERATION AND MAINTENANCE FINDINGS 4-1

Introduction 4-1
Circulating Water Inleakage 4-1
Air Inleakage 4-17
Other Operation and Maintenance Considerations 4-22

5 PERFORMANCE OF CONDENSER TUBE ALLOYS 5-1

Introduction 5-1
Copper-Base Alloys 5-12
Stainless Steels 5-31
Titanium 5-39
Summary of Performance 5-57

ix
Section Page

6 CONCLUSIONS 6-1
Performance of Condenser Tube Materials 6-1
Condenser Retubing Criteria 6-3
Impact of Condenser Leakage 6-4

7 RECOMMENDATIONS 7-1
Short-Term Work 7-1
Middle-Term Work 7-2
Long-Term Study 7-2

Appendix

A Bibliography by Number A-l

B Bibliography by Author B-l
C Keyword Listing C-l
D Mail Survey Form D-l
E Station Selection Matrix E-l
F Descriptions of Information Fields in Station F-l
Selection Matrix
G Field Survey Form G-l
H Field Survey Matrix Descriptors H-l
I Field Survey Matrix 1-1

LEGAL NOTICE

This report was prepared by Bechtel Corporation, as an account of work

sponsored by the Electric Power Research Institute, Inc. (EPRI).
Neither EPRI, members of EPRI, Bechtel, nor any person acting on behalf
of either: (a) makes any warranty or representation, express or implied,
with respect to the accuracy, completeness, or usefulness of the infor­
mation contained in this report, or that the use of any information,
apparatus, method, or process disclosed in this report may not infringe
privately 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.

X
 ILLUSTRATIONS

Figure Page

2-1 Steam Water Cycle 2-3

2-2 Single-Pass Surface Condenser 2-4
2-3 Single Pressure, Dual-Shell Condenser Design 2-6
2-4 Single Pressure, Three-Shell Condenser Design 2-6
2-5 Triple-Pressure Condenser with Long Tube Bundles 2-7
2-6 Triple-Pressure Condenser 2-8
2-7 Single-Pressure Condenser with Central Cold 2-8
Water Entrance
2-8 Heat Transfer Rate When F x x is 1.0 2-10
2-9 Correction to "U" for Cooling Water Inlet 2-10
Temperature
2-10 Flow Distribution over Area of the Exhaust 2-13
Connection
2-11 Tubesheet Configuration Designed for Steam 2-14
Entry All Around the Tube Bundle
2-12 Noncondensible Gas Collection in Single-Pass 2-15
Rectangular Surface Condenser
2-13 Noncondensible Gas Collection in Two-Pass 2-16
Rectangular Surface Condenser
2-14 Shrouded Air Removal Section for Cascaded 2-17
Removal of Noncondensibles
2-15 Location of Shrouded Air Removal Section of 2-18
Tube Bundle
2-16 Typical Tube-to-Tubesheet Joints 2-23
2-17 Typical Installation, Amertap Condenser Tube 2-25
Cleaning System
2-18 MAN System for On-Line Cleaning of Condenser 2-26
Tubes
2-19 MAN On-Line cleaning System for Condensers and 2-27
Tubular Heat Exchangers

xi
Figure Page

3-1 Typical Catalog File Card 3-3

3-2 Mail Survey Results of Condenser-Related Records 3-11
3- 3 Data Management Matrix 3- 21
4-1 Condenser Retubing Work Area 4- 3
4-2 Condenser Retubing-Tube Pulling Operation 4-3
4-3 Plants Surveyed that Achieve a Given Condenser 4-6
Water Leakage Detection-to-Localization Rate
4-4 Cation Conductivity Instrument Resin Columns 4-8
4-5 PWR Plants Surveyed that Detect Condenser Water 4-9
Leakage at a Given Bulk Condensate Impurity
Content
4-6 Fossil Plants Surveyed (>150 MWe) that Detect 4-10
Condenser Water Leakage at a Given Bulk Con­
densate Impurity Content
4-7 Air and Water Inleakage Rates for a Condenser 4-12
with 1.7 inch Hg Backpressure
4-8 Typical Tube Inserts (upper left) and Plugs 4-15
(lower right)
4-9 Typical Tube Insert Setup 4-16
4-10 Equilibrium Noncondensible Partial Pressure and 4-21
Equilibrium Condensate 0^ Content from Air In­
leakage vs Condenser Total Pressure and Sub­
cooling
4-11 Typical Circulating Water System for Twin Shell 4-24
Condenser and Open Cycle or Cooling Lake
4-12 Typical Characteristics of Low-Head Circulating 4-24
Water Pumps
4-13 Typical Characteristics of Medium Head Circulating 4-25
Water Pumps
4-14 Typical Cathodic Protection Electrodes 4-30
4- 15 Typical Tube Cleaning Devices 4- 30
5- 1 Failure Rate vs Probability Curves for 5- 15
Admiralty Brass in Freshwater Service.
Curves for Main-Body Service (M) (66
Tubesets) and the Air Removal Section
(A) (34 Tubesets) are Shown.
5-2 Failure Rate vs Probability Curves for 5-20
90-10 Cu-Ni in Freshwater (17 Tubesets)
and Seawater Service (56 Tubesets)

xii
Figure Page

5-3 Failure Rate vs Probability Curves for 5-23

90-10 Cu-Ni in Seawater Service, With
(29 Tubesets), and Without (29 Tubesets)
FeSO^ Additions.
5-4 Failure Rate vs Probability Curve for 5-25
70-30 Cu-Ni in Seawater (51 Tubesets,
of Which 6 are Main Body and 45 are
Air Removal).
5-5 Failure Rate vs Probability Curves for 5-27
70-30 Cu-Ni in Seawater With (24 Tubesets)
and Without (33 Tubesets) FeSO^ Additions.
5-6 Failure Rate vs Probability Curve for 5-28
Aluminum Brass in Seawater (37 Tubesets)
5-7 Failure Rate vs Probability Curves for 5-30
Aluminum Brass in Seawater Service With
(4 Tubesets) and Without (32 Tubesets)
FeSO^ Additions.
5-8 Failure Rate vs Probability Curve for 5-32
Aluminum Bronze in Seawater Service
(12 Tubesets)
5-9 Failure Rate vs Probability Curves for 5-35
Type 304 Stainless Steel in Freshwater
Service,- Curve A-A and M Tubesets (Guard
Plugs at One Station Deleted) (64 Tube-
sets); Curve B-A and M Tubesets (No Ex­
clusions) ; Curve C-I Tubesets (27 Tube-
sets) .
5-10 Failure Rate vs Probability Curve for 5-37
Type 304 Stainless Steel in Cooling
Tower Service (13 Tubesets). (Note:
As explained in the text, this curve
may be misleading).
5-11 Failure Rate vs Operating Hours and Failure 5-44
Rate vs Probability for Titanium Tubing
5-12 Failure Rate vs Operating Time for Main^-Body 5-61
Tubesets in Fresh Water
5-13 Failure Rate vs Operating Time for Main-Body 5-65
Tubesets in Seawater

xiii
 TABLES

Table Page

2- 1 Design Correction Factors for Condenser 2-12

Tubing Materials (F )
3- 1 Typical Computer Search History 3-2
3-2 Literature Search Keyword List 3-4
3-3 Station Selection Matrix - Age Distribution 3-9
by Fuel Type
3-4 Station Selection Matrix - Condenser Tube 3-9
Material by Cooling Water Category
3-5 Field Survey Stations - Age Distribution 3-12
by Fuel Type
3-6 Field Survey Stations - Condenser Tube 3-13
Material by Cooling Water Category
3-7 Field Survey Stations - Age Distribution by 3-14
Generator Size
3-8 Comparison of Station Selection Matrix with 3-15
Field Survey - Age Distribution by Fuel
Type
3-9 Comparison of Station Selection Matrix with 3-16
Field Survey - Condenser Tube Material by
Cooling Water Category
3-10 Comparison of Station Selection Matrix with 3-17
Field Survey - Distribution by Condenser
Manufacturer
5-1 Distribution of Tubesets by Material and Water 5-5
Type
5-2 Tubeset Failure Rate Calculation Sheet 5-8
5-3 Distribution of Tubesets with Known Failure 5-9
Rates by Material, Condenser Service, and
Water Type
5-4 Distribution of Number of Tubes and Condensing 5-10
Surface Area by Material and Water Type
5-5 Copper-Base Alloys, Chemical Composition and 5-13
Designation
5-6 Contributing Failure Mechanisms for Admiralty 5-14
Brass in Freshwater

xv
Table Page

5-7 Relationship Between the Shrouded-Unshrouded 5-17

Ratio, Rs, and Failure Rate
5-8 Contributing Failure Mechanisms for 90-10 Cu-Ni 5-22
in Seawater
5-9 Contributing Corrosion Failure Mechanism 5-24
for 70-30 Cu-Ni in Seawater
5-10 Contributing Corrosion Failure Mechanisms 5-29
for Aluminum Brass in Seawater
5-11 Contributing Failure Mechanisms for Aluminum 5-31
Bronze
5-12 Stainless Steel, Chemical Composition and 5-33
Designations
5-13 Contributing Failure Mechanisms for Type 304 5-38
Stainless Steel in Fresh Water
5-14 Maintenance and Cleaning Procedures Used for 5-40
Type 304 Stainless Steel Tubing
5-15 Field Survey Data, Titanium and Vibration 5-48
5-16 Field Survey Data, Titanium and Its Galvanic 5-56
Environment
5-17 Comparative Performance of Condenser Tube 5-59
Materials in Open Freshwater Cooling
Systems
5-18 Comparative Performance of Condenser Tube 5-63
Materials in Open Salt-Water Cooling Systems
5-19 Comparative Performance of Condenser Tube 5-68
Materials in Closed Cooling'-Water Systems

xvi
 SUMMARY

Condenser tube leakage has an impact on the operating cost, availability, and
useful service life of power generation facilities. Aware of this, EPRI con­
tracted with Bechtel Corporation to conduct a broad study of the factors that
affect the deterioration and subsequent leakage of condenser tubes. The results
of the study are given in this report.

The objective of the study was to provide the electric utility industry with
valuable guidelines for selecting and operating condensers and to highlight
areas that require increased research and development effort.

The study consisted of three broad tasks:

• A field survey of the condensers at 30 power stations with 76

generating units
• A literature search which resulted in a bibliography of 391
references pertinent to the study
• Analysis of the survey and literature data and the preparation
of the final report

Selection of the 30 stations for field survey was based on an analysis of 87

items of information relative to the design, operation, and performance of con­
densers at 87 power stations with 264 generating units. The information in this
selection matrix was compiled from the literature, government reports, Bechtel
files, and the results of a questionnaire mailed to the stations. The selection
process and the pertinent characteristics of the 30 stations surveyed as com­
pared to the profile for large, modern power generating units which was developed
by the selection matrix are described in Section 3.

A computer-based data management system was developed to enable efficient process­

ing of the hundreds of thousands of information items collected in the course of
the study.

xvii
Highlights of the findings of the study are summarized below.

The major impact of condenser leakage is the value of the power generation
lost while the unit is at part load or off line to repair the leaks. No
generally applicable cost figures could be determined because of the wide
variation in the costs assigned to lost generation by the various utility
systems dependent upon the unit down, the time at which it was down, the
system load at the time, etc. No demonstrated case of severe damage to any
components in the turbine cycle due to condenser leakage was found at the 30
stations surveyed.

Very small condenser leaks can be detected by analysis of the water chemistry
of the turbine cycle. Nearly all of the generating units surveyed have on line,
automatic analyzers for this purpose. The ability to locate the leak within the
condenser is not as sensitive as the available detection methods. However, most
plants are able to locate and repair any detectable leak. Hydrotesting the steam
side of the condenser was usually reported to be the most sensitive method of
leak localization. Another commonly used method is the application of either
plastic film or foam to the tubesheet of the evacuated condenser. There was no
agreement at the surveyed stations about which of these two methods is the more
sensitive. The impact of condenser leakage and the detection, location, and
repair of condenser leaks are discussed in Section 4.

Quantification of the performance and probable service life of condenser tube

materials was obtained by numerical analysis of the extensive condenser service
histories acquired by the study. The performance of common condenser tube ma­
terials in various environments is expressed numerically by a failure rate (FR),
which is the percentage of tubes plugged per 10,000 hours of operation. The
probability of attaining a specific failure rate is calculated from the data
base for each alloy, in each environment considered. From these probability
calculations, an estimate can be made of the anticipated service life the con­
denser tube material will have in the stated environment. The exposure environ­
ment is defined in terms of three circulating water categories and three con­
denser service sections. The circulating water categories are: 1) fresh water
once-through; 2) salt water; and 3) closed concentrating systems that use fresh
makeup water. The condenser service sections are: 1) impingement section;
2) main condensing section; and 3) air removal section.

xviii
It was found that, in the main condensing section of freshwater cooled condensers,
all of the commonly used materials have a high probability of lasting the plant
lifetime without retubing. In the air-removal section, however, admiralty tubes
do not give satisfactory service.

For the condensing section of saltwater cooled condensers, it was found that
of the materials surveyed only titanium has a high probability of lasting the
plant lifetime without retubing. Aluminum brass, aluminum bronze, and 90-10
Cu-Ni have a less than 50-percent probability of attaining a failure rate less
than 0.33, which corresponds to 40 years of service with 10 percent of the tubes
plugged.

Not enough information was obtained to allow a definitive evaluation of the

performance of the various condenser tube materials in closed cooling systems.
There is an indication that admiralty condenser tube performance is poorer in
closed systems than in open freshwater systems. Stainless steel condenser tubes
appear to give equally good service in both cooling modes.

From the accumulated data on condenser tube performance, an easily applied cri­
terion was derived to determine when a condenser is a candidate for retubing.
When the percentage of tubes plugged exceeds / H /100 (where H is the total
operating hours on the existing condenser tubes), a future retubing is likely.

Details of condenser tube performance are given in Section 5. A comparative

summary of the performance of the various condenser tube materials in the differ­
ent environments is given at the end of Section 5.

The conclusions of the study are briefly summarized in Section 6 and the recom­
mendations for future work are listed in Section 7.

xix
 Section 1

INTRODUCTION

STATEMENT OF THE PROBLEM

Condenser tube leakage has an impact on the operating cost, availability, and
useful service life of power generating facilities. Failure of condenser tubes
introduces circulating cooling water into the pure condensate and may thereby
contaminate the entire steam-water cycle. The degree of contamination is a
function of the rate of inleakage and the composition of the cooling water.

Two trends in power plant design and operation have further intensified the
impact of condenser tube leakage.

• The Environmental Protection Agency's 1974 Effluent Guidelines

and Standards for Steam Electric Power Stations (40 CFR 423)
have caused most new (and some existing) power stations to use
cooling towers or ponds for main condenser cooling in place of
once-through cooling. The resultant higher dissolved solids
in the cooling water and the generally higher pressure on the
condenser waterside magnify the contamination introduced by
condenser tube leaks.
• All U.S. manufacturers of PWR nuclear steam supply systems now
recommend the use of only volatile chemicals in secondary or
turbine cycle water treatment. Consequently, the introduction
of circulating cooling water into the condensate by condenser
inleakage will result in scale formation and/or increased
corrosion in the steam generators. Only minute quantities of
inleakage can be tolerated before the stringent limitations
on feedwater or steam generator water quality are exceeded
and the plant must be shut down.

Condensate demineralization can extend the rate of inleakage that can be tol­
erated before plant shutdown is required, but adds significantly to capital
and operating costs. There is also concern that, under certain circumstances,
condensate demineralizers will adversely affect water quality by sodium
contribution.

1-1
All of these considerations, together with the high cost of even a single
day's outage for modern large power plants, indicate how important it is
to take all reasonable measures to minimize condenser tube leaks.

The efforts of the industry to alleviate or eliminate condenser inleakage

have been hampered by the paucity and fragmentary nature of available infor­
mation on the frequency, size, and duration of condenser leaks; on the
anticipated service performance and life of various condenser tube materials
over the applicable range of exposure conditions; on the effect of condenser
design and construction parameters; on the sensitivity and accuracy of various
methods for the detection and location of condenser leaks; and on the best
procedures for repairing leaks with the minimum loss of power generation.

Aware of this problem, the Electric Power Research Institute (EPRI), in

November 1975, contracted with Bechtel Corporation to conduct a broad study
of the factors that affect the deterioration and subsequent leakage of
condenser tubes.

STUDY OBJECTIVE

As stated in the contract, the prime objective of this study is to "provide

the industry with valuable guidelines for selecting and operating condensers
and to highlight areas which require increased research and development
effort."

SCOPE OF WORK

The scope of work comprised three broad tasks:

• A field survey of existing power plant condenser systems

• A literature search and preparation of a bibliography of
references pertinent to this study
• Evaluation of the technical information so collected and prep­
aration of a final report with conclusions as well as recom­
mendations for further investigation.

To accomplish these tasks the work of the project was subdivided into the
following five phases.
1. Project Organization and Control

Upon award of contract the project was staffed with qualified, experienced
personnel. A detailed task flow diagram was constructed to delineate and
interrelate each task and subtask required to accomplish the prescribed work.
From this, a critical path schedule was derived and became the basis for schedule
and budget control throughout the project. Communication and reporting channels
between EPRI and Bechtel were defined. Progress reports and progress meetings
were scheduled at appropriate intervals.

2. Field Survey

Shortly after project inception, a 30-page field survey form was developed.
To guide the selection of stations for field survey that would constitute
a representative sample, a matrix of design and operating data was prepared
for 87 stations with 264 generating units. The input data for this selec­
tion matrix were compiled from a mail survey of the stations, published
reference material, and Bechtel records. The field survey covered 30 sta­
tions with 76 generating units. A computer data management system was estab­
lished to enable analysis and evaluation of the data obtained from the field,
mail, and literature surveys.

3. Literature Search

The initial effort was a search of the computerized portions of the Engineering
Index and the National Technical Information Service Index. This was followed
by a manual search of these and other indices and a cross-check of the refer­
ences listed in the discovered articles. Two listings of the bibliography of
391 references were prepared - alphabetically by author and numerically by
accession number. A keyword list was prepared for the articles in the bibliog­
raphy. Additionally, numerous unpublished articles and catalog items were
accumulated and indexed.

4. Evaluations and Conclusions

The data and information collected by the mail, field, and literature surveys
were evaluated and analyzed primarily by use of the computerized data manage­
ment system and prepared for inclusion in the final report. Conclusions and
recommendations were formulated.

1-3
5. Final Report

This report presents the results, findings, conclusions, and recommendations

that resulted from the study. A brief description of the contents of the
report is given in the following paragraphs to guide the reader.

Section 2 is a comprehensive review of the state of the art of main surface

condensers in power generation stations. The section describes the purpose
and configuration of condensers, discusses the more important design aspects
and parameters, and reviews the significant operating problems. No results
from the field survey program are included in this section.

Section 3 describes the methodology employed to accomplish the scope of work.

The technique and results of the literature search are described. The mechanics
for selection of the stations for the field survey are described, and the vali­
dity of the selected sample is evaluated. The conduct of the field survey and
the data sought are discussed. The data management system developed to process
and analyze the collected data is explained. No results from the field survey
program are given.

Sections 4 and 5 present the results and findings of the mail and field survey.
Section 4 deals with the occurrence of condenser leaks; their impact on power
plant operation and equipment; methods for detection, and location, and repair
of condenser leaks; condenser tube cleaning methods; air inleakage and con­
denser maintenance; and station records related to the condenser.

All data related to the performance of various condenser tube materials in

different operating environments are presented in Section 5. The calculated
failure rates and probable service life of each material in specific environ­
ments are given. The section discusses copper alloys, stainless steel alloys,
and titanium. There is a concluding discussion of considerations for deciding
whether to retube a failing condenser.

Section 6 draws together the significant findings and conclusions that resulted
from the study.

Section 7 formulates the proposed programs for further investigation and de­
velopment that were derived from the work of this study.

1-4
 Section 2

CONDENSER FUNCTION, DESIGN, AND OPERATION

This Section describes the state of the art for modern, power plant condensers.
The major topics of discussion are:

• Purpose of the condenser in the steam power plant cycle

• Terminology and functions of the major condenser components
• Typical condenser layout arrangements
• Some pertinent features of condenser design

PURPOSE OF THE CONDENSER

The major function of the main condenser is to transfer the heat rejected from
the steam cycle to a receiving body of water or to the atmosphere when a
cooling tower is employed. The heat transferred is the latent heat of
vaporization of the turbine exhaust steam. The rejected heat comprises 60 to
75 percent of the heat generated in the fuel. The total quantity of heat
transferred is the "condenser duty" which is usually expressed in Btu/hr.

The condenser also functions as a barrier between the cooling water and
the condensate. Modern, high-pressure boilers require high purity feedwater.
Less than 1 ppm of total dissolved solids (TDS) is desirable. Cooling
waters range in TDS from less than 100 ppm for high quality fresh water
used once-through, up to 35,400 ppm for normal seawater used once-through.
Any intrusion of cooling water into the condensate is undesirable. The
quantity of inleakage that can be tolerated varies, depending primarily
on the TDS of the cooling water. Of the constituents that make up the
TDS of the cooling water, some such as silica and those contributing to hardness
and alkalinity are more significant as pollutants to the feedwater than are
others such as sulfate. Inleakage of cooling water can result in scale deposi­
tion in the feedwater heaters and boiler, and in increased corrosion throughout
the "turbine cycle."

2-1
The condenser also serves to remove noncondensible gases from the turbine cycle
and, in some cases, to reheat the condensate above the saturation temperature of
the condenser.

The function of the condenser in the thermal power cycle is illustrated by

Figure 2-1. This figure is a schematic of the steam water cycle for a typical
fossil-fueled power plant.

While the turbine cycle for BWR and PWR nuclear power plants is different than
that for fossil-fueled power plants, the essential function of the condenser is
the same. One difference is that the condenser in a nuclear plant has a con­
siderably higher duty per unit of power generated because of the lower thermo­
dynamic efficiency of current nuclear plants.

CONDENSER CONFIGURATION

A typical configuration of a condenser shell is shown in Figure 2-2. This is an

isometric cutaway view of a single pass (of the cooling water) condenser with a
divided water box. The illustrated condenser has two bundles of tubes, each
with its own air removal section. Cooling water flows in parallel through the
two bundles. It is conventional to introduce the exhaust steam from the LP
turbine into the top of the condenser as shown at point (12) in Figure 2-2.
The cross-sectional area of the steam path expands from the turbine flange, or
exhaust connection (O) to the top of the tubing bundle (ll) . The transition
section is the exhaust neck or steam dome. The outer few rows of tubes on the
top and sides of the tube bundle are the impingement section (ll) . The exhaust
steam flows throughout the tube bundle condensing section (16) via open lanes
in the tube layout pattern. The noncondensible gases are swept along to the air
removal section (s) whence they are withdrawn by vacuum pumps or steam jet ejec­
tors. The condenser tubes are affixed by rolling or welding to the tubesheets
(9), and supported along their length by support sheets (19) . Other waste steam
enters the condenser at (?) > which is the feed pump drive turbine exhaust,
and at (l5) which shows some of the many drain and dump connections. Conden­
sate falls to the hotwell (10) and is withdrawn via the condensate outlet
(20) by the condensate pumps. The cooling water enters at connection (7) and
is distributed across the tubesheet: by the "cold end" water box • Noncon­
densible gas accumulating in the water box is removed via connection (4). The

2-2
to
I
OJ

Figure 2-1. Steam Water Cycle

 ,11

1. EXTRACTION CONNECTION TO 11. TUBE BUNDLE (IMPINGEMENT)

EXTERNAL HEATER
12. TURBINE EXHAUST HOOD
2. DIAPHRAGM CONNECTION BETWEEN CONNECTION
TUBE SHEET AND SHELL
13. EXPANSION JOINT TO TURBINE
3. FEED PUMP TURBINE DRIVE EXHAUST EXHAUST HOOD

4. NONCONDENSIBLE REMOVAL 14. F.W. HEATERS IN CONDENSER STEAM

CONNECTION DOME
5. MANHOLE 15. DUMPS AND DRAINS
6. WATER BOX 16. TUBE BUNDLE (CONDENSING)

7. COOLING WATER CONNECTION (IN) 17. COOLING WATER CONNECTION (OUT)

8. TUBE BUNDLE (NONCONDENSIBLE 18. HOTWELL LEVEL

REMOVAL SECTION)
19. TUBE-SUPPORT SHEETS
9. TUBE SHEET
20. CONDENSATE PUMP CONNECTION
10. HOT WELL
21. COLD END BAFFLE IN
NONCONDENSABLE SECTION

Figure 2-2. Single Pass Surface Condenser

Source: Westinghouse

2-4
cooling water flows inside the tubes to the "warm end" water box and exits at
connection (tt) . Access to the interior of the water boxes and to the steam
space of the condenser is provided by appropriate manholes (T). Differential
thermal expansion of the components is accommodated by the tubesheet-to-shell
diaphragm (?) and turbine exhaust hood expansion joint (13) . Some of the low
pressure feedwater heaters are often located in the condenser neck (Q) .

Condenser Layout Arrangement

Various condenser layout arrangements are made to accommodate numerous factors

in the overall power plant design. Some of these considerations are the effi­
cient layout of the turbine floor area, the lowest overall evaluated power
generation cost for the plant, etc. Throughout this report, the following gen­
erally accepted terminology is used. Each generating unit has only one conden­
ser. This condenser may consist of one to three separate shells. Each shell
may contain from one to four tube bundles. The cooling water flow may be single
pass, i.e., parallel flow through all bundles, or multipass, i.e., series flow
through two or more bundles located in the same shell. The pressure in steam
space may be the same in all shells, i.e., single pressure (usually accompanied
by single-pass cooling water flow), or the steam side pressure may be slightly
different in each shell, i.e., multipressure (usually accompanied by single-pass
cooling water flow in series through the shells).

Examples of various condenser layout arrangements are illustrated in Figures 2-3

through 2-7. The condensers shown in Figures 2-3 and 2-4 are transverse arrange­
ments in which the condenser tubes are perpendicular to the turbine shaft:

• Figure 2-3 shows a two-shell, single-pressure, single-pass

condenser with four tube bundles.
• Figure 2-4 shows a three-shell, single-pressure, single-pass
condenser with six tube bundles. Separate small condensers
are provided for the exhaust steam from the auxiliary turbines
which drive the feedwater pumps.

Figures 2-5 through 2-7 show longitudinal condenser layouts wherein the conden­
ser tubes are parallel to the turbine shaft:

Figure 2-5 shows a single shell, divided by intermediate tube

sheets to form a three-pressure condenser with a single-pass,
series flow cooling water path. There are two parallel tube
bundles within each pressure zone. This results in very long
condenser tubes.

2-5
Figure 2-3 Single Pressure, Dual-Shell
Condenser Design

Figure 2-4. Single Pressure, Three-Shell

Condenser Design

2-6
 824,000 SO. FT.

Figure 2-5. Triple-Pressure Condenser with

Long Tube Bundles

• Figure 2-6 shows a single-shell, three-pressure condenser

similar to that shown in Figure 2-5. An intermediate waterbox
is provided in the center to reduce the length of the conden­
ser tubes.

• Figure 2-7 shows a single-shell, single-pressure condenser

with three exhaust compartments. The central cooling water
inlet boxes provide for single-pass flow through the four
parallel tube bundles.

2-7
r
i-*
-

Figure 2-6. Triple-Pressure Condenser

LP HEATERS
LP HEATERS

STEAM DUMP

FP EXHAUST

618,150 [Link]. EPEXHAUST

Figure 2-7. Single-Pressure Condenser with

Central Cold Water Entrance

2-8
CONDENSER DESIGN CONSIDERATIONS

The thermodynamic and structural design of large, modern power plant condensers
is a complex undertaking. A rigorous discussion is outside the scope of this
report. However, some of the principles of condenser design will be discussed
in a simplified approach which will be adequate for the context of this report.

General

The essentials of the heat transfer relationships in condenser design can be

expressed by simplified, basic equations.

The total surface area of exposed tubing, S, required in the condenser is:

c _ DutY
0
where,
S = surface area, square feet
U = heat transfer rate Btu per square foot of surface per
°F of log mean temperature difference
0^ = °F log mean temperature difference

and,
T - T„

s o
where,
T^ = cooling water outlet temperature, °F
T^ = cooling water inlet temperature, °F
T = temperature °F of saturated steam at the
absolute pressure of the condenser
and,
U = F1 • F2 - F3 • C • /~V
where.
F^ = cooling water temperature factor
F2 = tube material and wall thickness factor
F^ = tube cleanliness factor
V = average water velocity in tube, ft/sec
C = condenser tube heat transfer constant

Values for U when F^ x F2 x F^ = 1.0 are shown in Figure 2-8 for various tube
diameters (C values) and water velocities. These values are for new, clean,
bright, and oxide-free tubes. The value for F^, the factor for inlet water tem­
perature is obtained from Figure 2-9. The value for F^, the factor for tube

2-9
 VELOCITY FEET PER SECOND

Figure 2-8. Heat Transfer Rate When

F^ x x is 1.0
H
—

:m;
DESIGN CORRECTION FACTOR (

ip i:ir iM Mi
Irj ■- r iii ;M 1 MM- n-L, :.n!
'C O - -

4-i.
$ Hit t: Mi
Mr lit iM ?rr Mr
1 W 1 •~TTt ! i' ii!i it; i HiliH
$
1 tiMit '!M M:l 1 iM
(if MM
ill
i
irti tlj- .!]!■] .illj 4i Mi u iM :.jh tii |ti. si: l tt
4:! f:ii iH i3i
T‘
1 i-i: 1 'M •j IM tjt! tin ;:t 1 i 1 it. i.M 43 iiij
it fij tt
'!1
o

'tin 1L !;l; Mi if lilt t:j+; iiji tl| tt M; It lr ii t

ft!
1 in 1 !:l; jM EUP .pt !1 • ipi ii | '4 Mi iiii Mi 1 4- it! H 4 HL t
B

44 4:: Lt
O

Ifir sS i'i; 'it M’ ft it; Mr it:

p

•ni r\\ ijM jii! Mi it: M: !M M Mt M* :rt

4 Mi
1 ill) 1 $ M-i iH !i|l Hi Mi Ht- m! !t: Hr tt 4 4 fl ;t ijil i-l; Iiii Ei
1
-J

44 Mi
P

:4: 1: tilj iri |;m 41

Mi
Mp Mi itti tt M-;
1 M': 1' iM ..IM, ;:M t
1 fl 1 1 1 iijt f IT;: It | it it Mi* lit
1
O'

[Link]
P

i^r .iir! |lii ill :H IH it

iU 1] !M i i| |iii MMM :M !m
;i!i Hi 1 Ifi M: • i *' Ml Ml lr! 1 1¥ 1 iM 1 W1 1 m; Mi I
>
n

MM

JO 40 50 60 70 80 90 l<
TEMPERATURE OF INLET WATER DEGREES F

Figure 2-9. Correction to "U" for

Cooling Water Inlet Temperature

Source: HEI, Standards for Steam Surface Condensers, Sixth Edition.

2-10
material and wall thickness is obtained from Table 2-1. The value of is the
design cleanliness factor assumed and it is a proportion of the performance of
new, clean, and bright tubes. Conventionally, is taken as 0.85 for copper
alloy tubes and as 0.90 for stainless alloy or titanium tubes. The cleanliness
factor is increased by 0.05 when a continuous tube cleaning device (e.g., Amer-
tap) is used.

The required cooling water flow, Q is approximately:

_ Duty____ Q is gpm
^ 500(T - T ) Duty is Btu/hr

The optimization of the condenser design parameters is a complex process. To

select the best combination of surface area; tube diameter, gauge and length;
temperature rise; etc. involves many economic and technical tradeoffs. Conden-
Sgr optimization is usually done with the assistance of an appropriate computer
prograift. In the selection of the tube material consideration must be given first
to its servic£3-bility in the particular environment, and then to its heat trans­
fer and mechanical properties. The ratio of surface area to turbine rating var­
ied from 0.3 sq ft/KW to 1.05 sq ft/KW at the 30 stations surveyed by this study.

Steam-Side Considerations

Steam Velocity. The expanded st-eam exhausts from the low pressure turbine at
high velocity and with a considerable moisture, content. The exhaust loading
varies from 7,500 Ib/hr/sq ft of area at the exha'jst connection to 15,000 lb/
hr/sq ft. When the absolute pressure of the condenser is low (due to low
cooling water temperature) the steam exhausts from the turbine at or near sonic
velocities. The distribution of the steam flow is not uniform over the exhaust
connection area. There is a concentration of flow (hied1 velocity region) ad­
jacent to the bearing end of the turbine. This is illustrated in Figure 2-10
which is a plot of a free flow air test pattern for a turbine of current design.
A transition section (exhaust neck or steam dome) is needed to reduce its velo­
city. The tubesheet configuration must provide proper lanes to distribute the
steam to within the tube bundle, and to transport the noncondensible gases to
their collector. A typical tubesheet configuration is shown in Figure 2-11.

Removal of Noncondensible Gases. The noncondensible gases consist primarily

of air which is drawn into the evacuated condenser and its subatmospheric
appurtenances through leaks at packing glands, welds, threaded joints.

2-11
 Table 2-1*

DESIGN CORRECTION FACTORS FOR CONDENSER TUBING MATERIALS (F )

TUBE WALL GAUGE - BWG

TUBE MATERIALS
24 22 20 18 16 14 12

Admiralty Metal 1.06 1.04 1.02 1.00 0.96 0.92 0.87

Arsenical Copper 1.06 1.04 1.02 1.00 0.96 0.92 0.87
Aluminum 1.06 1.04 1.02 1.00 0.96 0.92 0.87
Aluminum Brass 1.03 1.02 1.00 0.97 0.94 0.90 0.84
Aluminum Bronze 1.03 1.02 1.00 0.97 0.94 0.90 0.84
Muntz Metal 1.03 1.02 1.00 0.97 0.94 0.90 0.84
90-10 Cu-Ni 0. 99 0.97 0.94 0.90 0.85 0.80 0.74
70-30 Cu-Ni 0.93 0.90 0.87 0.82 0.77 0.71 0.64
Cold-Rolled Low
Carbon Steel 1.00 0.98 0.95 0.91 0.86 0.80, ''6T74
Stainless Steels
Type 410/430 0.88 0.85 0.82 0.76 o.?u' 0.65 0.59
Type 304/316 0.83 0.79 0. 75 0.69 0.63 0.56 0.49
Type 329 0.78 0.76 0.74 0.6S 0.65 0.60 0.54
Titanium (tentative) 0.85 0.81 0.77 0.71 —
—— —

*Source: HEI, Standards for Steam Surface Condensers, Sixth Edition.

2-12
 TTTrTTrTnj////////1////////////////////////n TpTTTnT//////////TnYrrrrTTfTiln/l/l///////////if 7

15'-6"

7'-9"

TRANSVERSE CENTER PLANE

DOUBLE FLOW - DOWN DISCHARGE - 38" OR 43" LBS - HERZOG HOODS

BASED ON AIR TEST TRAVERSES FROM SET 231
REPLACES 327 HA 481

NUMBERS ON CURVES DENOTE THE RATIO OF LOCAL FLOW

PER UNIT AREA TO AVERAGE FLOW PER UNIT AREA

Figure 2-10 Flow Distribution over Area of the

Exhaust Connection. Source: General Electric

2-13
 Figure 2-11. Tubesheet Configuration Designed for
Steam Entry All Around the Tube Bundle

instrument connections, etc. In the steam from boiling water reactors there
is a significant concentration of hydrogen and oxygen due to the radiolytic
decomposition of water within the nuclear fuel core. In fossil and PWR plants
which employ volatile feedwater treatment additives the noncondensibles include
a small volume of ammonia and the decomposition products of hydrazine and, pos­
sibly, of other volatile amines. The noncondensible gases must be promptly
evacuated from the condenser to prevent their rendering a part of the heat
transfer surface ineffective due to air binding.

One method for removal of the noncondensibles is shown in Figure 2-12 as

applied to a single-pass condenser, and again in Figure 2-13 as applied to a
two-pass condenser. This method is also illustrated in Figure 2-2. The core
collector draws the noncondensibles, together with some water vapor, from be­
tween each pair of tube support plates and transports the mixture to the cold
end of the condenser. Here the mixture is passed over the coldest tubes, to
reduce the water vapor concentration to the minimum possible, and then with­
drawn by the air removal system. The noncondensible core collector is not
shrouded. The condensate formed on the tubes above flows down through the air
removal zone. Any ammonia concentrated in the air removal zone can be absorbed
by this condensate up to equilibrium limitations. The quantity of ammonia re­
moved with the noncondensible gases is at the resulting equilibrium minimum.

2-14
Core Air-OfFtake
Features
i) Non-condensables are removed from the center of tube banks providing -
Short steam flow paths
Low pressure losses
• 2.) Core air-offtake pipes have a varying number of orifices in each bay to
insure that each bay can capably handle the volume of noncondensables
associated with the condensing action, thus balancing the venting properly
along their entire lengths.
3) Non-condensables are removed by first passing them over a few rows of
tubes to subcool the air and vapor mixture adequately - then transporting
them out of the condenser through pipes located in the water boxes, and
thence to the air-removal system.
4) Each bundle is individually vented to assure proper removal of non­
condensables.

Figure 2-12 Noncondensible Gas Collection in

Single-Pass Rectangular Surface
Condenser. Source: Westinghouse

2-15
 CORE AIR OFFTAKE SYSTEM
TWO PASS

Core Air-OfFtake
Features
x) Non-condensables are removed from the center of tube banks providing -
Short steam ilow paths
Low pressure losses
z) Core air-offtake pipes have a varying number of orifices in each bay to
insure that each bay can capably handle the volume of noncondensables
associated with the condensing action, thus balancing the venting properly
along their entire lengths.
3) Non-condensables are removed by first passing them over a few rows of
tubes to subcool the air and vapor mixture adequately - then transporting
them out of the condenser through pipes located in the water boxes, and
thence to the air-removal system.
4) Each bundle is individually vented to assure proper removal of non­
condensables.

Figure 2-13. Noncondensible Gas Collection in

Two-Pass Rectangular Surface
Condenser. Source: Westinghouse

2-16
Another method for removal of the noncondensibles is shown in Figure 2-14. In
this design, the air removal section is shrouded. The placement of the shrouded
air removal section within the tube bundle is shown in Figure 2-15. The mixture
of noncondensibles and water vapor collected between tube support plates is cas­
caded to the cold end of the condenser. The water vapor content is minimized by
drawing the mixture over the coldest tubes just prior to its withdrawal by the
air removal system. The shroud prevents "washing" of the collected noncondensi­
bles by the condensate from above. Soluble gases such as ammonia are removed
from the condenser more effectively as compared to the open air removal section
described above.

The design capacity of the air removal (venting) system is given in the Heat
Exchange Institute "Standards for Steam Surface Condensers," 6th Edition. The
capacity provided is a function of the number of turbine exhaust openings, the
steam flow per exhaust opening and the concentration of dissolved oxygen desired
in the effluent condensate. Either steam jet ejectors or vacuum pumps are used
to evacuate the noncondensible gases from the condenser.

WARM
END

COLD
END

Figure 2-14. Shrouded Air Removal Section for

Cascaded Removal of Noncondensibles.
Source: Ingersoll-Rand

2-17
p OOO O O O O O oon O 0,0 O OOO O O O O O OOO o
NORTH END

£ fe'op^cbcoro'o0o^^:o'o®o”o0o0d'o'cyo'4J V/.

O °&i-

: , : 1
o'/; ;,■.•/// II' r 1 , ii; Si • ■ •.•>. v.o'^‘J0

OOOO O OO C.O OOO OOO O O OOO O O O O O OOO

Figure 2-15. Location of Shrouded Air

Removal Section in Tube Bundle

2-18
Water-Side Considerations

Cooling Mode. Whether the cooling mode is open cycle or closed cycle has a very
significant impact on many water-side design considerations. Open cycle is
once-through cooling with either fresh or salt water. Closed cycle is any con­
centrating circulating water system and includes cooling towers, cooling ponds,
and spray ponds or canals which recycle the cooling water. The cooling mode
directly or indirectly affects the cooling water inlet temperature, the cooling
water temperature range, the exposed condensing surface area, the waterbox pres­
sure, the stress on the tube-to-tubesheet joint, related structural and mechanical
design requirements, etc. These impacts are discussed in more detail under the
individual headings which follow.

In addition to the above impacts on condenser design, the cooling mode can also
affect the efficiency of the generating plant. Plants with open cooling systems
can take advantage of the generally lower cooling water temperature and the re­
sulting lower absolute pressure in the condenser to justify economically a tur­
bine with a larger back end. This usually results in a lower heat rate for the
plant. In plants with closed cooling water systems, the turbine must exhaust
against a higher average condenser backpressure. This may result in the choice
of a smaller back end and the plant heat rate will be higher.

As brought out in Section 3, a substantial majority of the power plants now in

operation in the United States use the open cooling mode. However, under the
present regulations for effluents from steam electric power generating stations
(40 CFR 423), all power plants which were first placed in service after January 1,
1974 (after January 1, 1970 for plants larger than 500 MWe) in effect must employ
a closed cooling water system unless an exception is granted within the provisions
of 40 CFR 122 (Thermal Discharges).

Temperature Range. The cooling water temperature range, T , is defined as:

Both the inlet cooling water temperature (T ) and the temperature range (T ) in-
fluence strongly the condenser's design and performance. As shown earlier in
this Section, the required exposed condensing surface area is inversely propor­
tional to the log mean temperature difference which, in turn, is a function of
the temperature range. Also, the quantity of circulating water required is
directly proportional to T .

2-19
As a practical matter, both the average inlet temperature and average temperature
range are lower for open systems than for closed systems. The temperature of
natural surface water bodies is usually well below the wet bulb temperature of
the ambient air over a considerable portion of the year. Economic design con­
siderations result in an "approach" for closed evaporative cooling systems in
the range of 8 to 15°F. The approach is the difference between the temperature
of the cooled water (inlet water to the condenser) and the ambient air wet bulb
temperature.

In the existing open cooling systems, the temperature range (T ) generally is from
15 to 20°F. This range is commonly limited by the impact of the thermal dis­
charges on the receiving water. In closed cooling water systems, the temperature
range is usually from 30 to 40°F. The temperature range for closed cooling sys­
tems is generally determined by economic conditions.

Tube Velocity. As shown above the heat transfer rate for a given condenser tube
is proportional to the square root of the average cooling water velocity. From
just this consideration, then, high velocities are desirable in order to reduce
the required condensing surface area. On the other hand, higher velocities in­
crease the friction loss and therefore the pumping energy. Also, higher veloci­
ties can cause severe erosion of the condenser tube material. The selected
velocity must balance these considerations, along with many others, in the opti­
mization of the condenser design. In practice, a velocity of 7 ft/sec is used
almost exclusively with copper alloy tubes. With stainless steel tubes, a
velocity of 7 to 8 ft/sec is usual. At every one of the 30 stations surveyed in
the field the water velocity was a nominal 7 ft/sec despite variations in tube
material and circulating water composition.

Waterboxes. The principal function of the waterbox is to evenly distribute the

incoming water across the tubesheet. Ideally, the velocity of approach would be
constant at all points. In nearly all condensers the waterbox acts as a 90-degree
elbow to turn the incoming water into the tubesheet. When waterbox pressure is
only slightly above atmospheric (or even below atmospheric in some syphon leg
circulating water systems), previously dissolved air is released. The released
air must be removed from the waterbox to prevent air binding of the condenser
tubes on the water side. Air release, together with improper inlet water distri­
bution, can induce inlet-end erosion of the condenser tubes. In condensers that
have experienced inlet-end erosion, open-end plastic or metal tubes have been
inserted 6 to 12 inches into the condenser tube to alleviate the problem.

2-20
To combat cavitation, erosion, and corrosion, waterboxes usually have a pro­
tective coating, a cathodic protection system, or a corrosion-resistant metal
lining. Common protective coatings include epoxies and other plastics, fiber­
glass, and rubber. Cathodic protection systems may consist of either sacrificial
anodes affixed to the waterbox or impressed current systems. The cooling water
must have a low electrical resistance for cathodic protection to be economical.

In open cooling systems, the cooling water body is almost always at an elevation
below that of the condenser waterboxes. Hence, the pressure in the waterboxes
and the condenser tubes seldom exceeds a few psig and can be subatmospheric. In
a closed cooling system which uses a cooling tower, the bottom of the cooling
tower is commonly at a higher elevation than the waterboxes. Also, the cooling
water must be lifted an additional 30 feet or more to the top of the cooling
tower fill. There results a waterbox pressure of 30 psig or more depending on
the relative elevation of the cooling tower to the condenser and the pump charac­
teristics. Due to the higher pressure in a closed system, a given size of open­
ing in a condenser tube will pass more circulating water into the condensate than
would an open system. The increase in inleakage flow varies as the square root
of the increase in pressure.

Tube-to-Tubesheet Joints. A tight tube-to-tubesheet joint obviously is mandatory

to prevent circulating water inleakage. The tube joint also serves an important
structural function. The joint must resist the pull-apart forces which exist
when the condenser is hydrotested on the steam side with empty waterboxes to
detect joint or tube leaks. The joint must also resist the tension or com­
pression forces that can occur under other conditions. Because of the holes in
the tubesheet, the net forces on the waterbox can place tensile or compressive
forces on the tubes. The nature and magnitude of net force depends primarily
upon the waterbox pressure, the steam space pressure, and the fraction of the
tubesheet that is drilled out. Severe tensile or pull-apart forces can exist in
an unevacuated condenser with high waterbox pressure, e.g., closed cooling tower
circulating water system.

Common tube-to-tubesheet joints are the following:

• Plain rolled joint

• Grooved rolled joint (single or double grooved)
• Welded joint (seal weld or strength weld)

2-21
 • Grooved rolled joint with overlapping internal grooves that
interconnect all tubes
• Packed joint which allows tube expansion when the tubesheet-
to-condenser shell joint is rigid
• Double tubesheet joints - the tube is rolled into two closely
spaced tubesheets at either end.

Most of these joints are illustrated in Figure 2-16. In existing power plants,
the first two types of joints are by far the most common. It is beyond the scope
of this report to analyze the technical and economic aspects of the several join­
ing systems.

MAINTENANCE OF TUBE CLEANLINESS

While in service, condenser tubes develop deposits of foulants which reduce the
effective heat transfer rate. The foulants may be metallic oxides, mineral
scales, or biomasses.

All condenser tube materials, on exposure to circulating water and air, develop
an adherent film of oxides of the metals in the alloy. This oxide film protects
the base metal against further corrosion and normally does not form to an extent
that significantly reduces design heat transfer rate.

Mineral scales such as calcium carbonate, calcium sulfate, aluminum silicate, etc,
are very rare in open cooling systems. In closed cooling systems, such scales can
occur when proper circulating water chemistry is not maintained. Mineral scales
cause a severe reduction in heat transfer rate and an attendant increase in con­
denser pressure. When scale does occur, its removal may require mechanical or
chemical cleaning.

Biomasses, commonly composed of filamentous algae or slime forming organisms, are

the most prevalent foulant, in both open and closed cooling systems.

Control of tube fouling is accomplished by application of chemical biocides or

by mechanical cleaning. Mechanical cleaning devices are available which can be
used with the condenser in service or off-line.

2-22
WATER FLOW

TUBE

OVERLAPPING
GROOVES
TUBE SHEET

(C) GROOVED AND ROLLED - WITH (D) PACKED JOINT

PRESSURIZED OVERLAPPING GROOVES
TO SEAL THE C.W.

Figure 2-16. Typical Tube-to-Tubesheet Joints

2-23
Biocides

While a number of suitable biocides are available, chlorine is the unanimous

choice at power generating stations. Chlorine is selected because of its bio­
cidal effectiveness, its lower cost, and its availability. Chlorination re­
quirements vary seasonally since bioactivity (especially of photosynthetic algae)
is affected by sunlight. Chlorination practices vary within the industry but a
typical plant may chlorinate once or twice a day in the winter and three or four
times a day in the summer. Chlorine is applied at a rate that will produce a
chlorine residual of 0.5 to 1.0 ppm in the water leaving the condenser. Each
chlorine application may last 15 to 60 minutes.

Mechanical Cleaning Devices

On-Line. The two available systems for on-line cleaning are the Amertap system
and the MAN system.

In the Amertap system, sponge rubber balls are introduced into the inlet water-
box. The neutral buoyant density of the balls enables a random distribution of
balls over the entire tubesheet. The balls are driven through the tubes by the
circulating water. The balls are slightly larger in diameter than the tube ID;
hence a cleaning of the tubes results. The balls are retrieved in a catcher
device in the outlet circulating water line. After automatic classification to
eliminate defective balls, the balls are recycled by a pump to the inlet water-
box. Figure 2-17 illustrates the Amertap system.

The MAN system consists of a stiff brush retained in each condenser tube by suit­
able end-cages. The circulating water piping is arranged to allow periodic rever­
sal of flow. Each flow reversal drives the brush through its respective tube.
The MAN system is shown in Figures 2-18 and 2-19.

Off-Line. Over the years, condenser tubes have been cleaned by manual rodding
with brushes or plugs. One currently used system for cleaning the tubes consists
of inserting a specially designed scraper in the tube and propelling it rapidly
through the tube by a hydraulically operated gun.

2-24
INJECTION NOZZLE

BALL VALVE

DISTRIBUTORS

y LHYOROFOtL

CDOLING WATER
OUTLET

Figure 2-17. Typical Installation, Amertap Condenser Tube Cleaning System

2-25
 normal c. w. flow direction

c.w. inlet c.w. outlet

Figure 2-18. MAN System for On-Line Cleaning of Condenser Tubes

2-26
Figure 2-19 MAN On-Line Cleaning System for Condensers
and Tubular Heat Exchangers

2-27
 Section 3

METHODOLOGY OF THE PRESENT STUDY

INTRODUCTION

This section describes the methodology employed to accomplish the major work
items, which will be discussed in the following order: 1) Literature Survey;
2) Matrix for Selection of Field Survey Stations and Mail Survey; 3) Field
Survey; and 4) Data Management System.

LITERATURE SURVEY

Mechanics

The initial approach to the literature survey was a search conducted by Bechtel
central library personnel of the computerized portions of the Engineering Index
(Compendex), which contains entries dated from 1970 to the present, and of the
National Technical Information Service Index, which dates from 1964. The first
step was to determine the number of titles listed under various keywords or
descriptors and appropriate sequential combinations of the descriptors. A
sample search history is shown in Table 3-1. From these listings, abstracts
were requested for approximately 650 references in the Engineering Index and
85 in the NTIS Index.

Simultaneously, a search was made for abstracts of pertinent articles in the

computerized reference file of the Data Center of the Copper Development
Association. CDA kindly furnished more than 120 abstracts.

After all of these abstracts were reviewed, Bechtel library personnel were
asked to obtain copies of some 350 articles that appeared to be relevant to
the study but which were not yet in the project files.

Next, a manual search was made of the Engineering Index from 1960 to 1970.
This search employed those keyword descriptors found most productive during the
computer search. Some 110 additional articles were discovered that were deemed
pertinent and copies of those references were sought.

3-1
 Table 3-1
TYPICAL COMPUTER SEARCH HISTORY

****m****mm********************************* ****************
* *
* *
* *
* Lockheed Retrieval Services *
* *
* *
* *
************************************************************

Dser 965 Date:01/16/76 Tin[Link] File: 8

Search History Print Summary

Set Itess Description No. File Accn/Set Fat Item-range
1 9025 COPPER 1 8 34 5 1-409
2 12109 ALUHISOH 2 8 60 5 1-240
3 3554 STAINLESS{«)STEEL
4 3836 TITANIOB
5 75 ADHIRALTY
6 8 HONTZ
7 920 BRASS
8 511 BRONZE
9 26194 1-8/OR
10 8368 CORROSION
11 10372 TUBE?
12 1324 EROSION
13 2801 9AND10
14 763 SAND11
15 133 9AND12
16 2261 INSPECTION
17 4579 VIBRATION
18 26774 TESTING
19 2081 CA=616
20 42370 10OH12OR16OR17OR18OR19
21 1681 20AND11
22 1681 20AND11
23 4905 130R140R150R22
24 0 WATER BOX
25 0 WATERBOX
26 0 WATER BOXES
27 0 WATERBOXES
28 119 TOBE (F)SHEET 45 10077 35-44/OR
29 11 TOBESHEET 46 0 CON DENS
30 35 TOBE(F)SHEETS 47 62 FEEDWATER(F)HEATER S
31 6 TOBESHEETS 48 803 CONDENSER?
32 153 28-31/OR 49 304 CONDENSING
33 271 9AND11AND20 50 1047 480R49
34 409 320R33 51 190 50AND11
35 2803 CA=613 52 246 45AND50
3759 CA—614 53 40 47AND45
2706 NUCLEAR (F)POWER(F)PLANTS 54 280 520R53
6522 POWER <F)PLANTS 55 439 510R54
39 117 FOSSIL(F)PLANTS 56 802 340R55
40 82 FOSSIL(F)PLANT 57 6394 TUBES
41 1168 POWER (F)STATION 58 301 55N0T57
42 1032 POWER(F)STATIONS 59 132 57AND50
43 315 CENTRAL (FJ STATION 60 392 540R59
44 180 CENTRAL(F)STATIONS 61 36530 STUDY

3-2
Additional articles were obtained during the field surveys and meetings with
people in the industry.

As articles were received, each was given an accession number assigned in

approximate order of receipt. For each article a catalog card was prepared
which included the authors' names, title of the article, its reference citation,
an abstract of the article, and a listing of its keyword descriptors. A
typical file card is shown in Figure 3-1. The list of keywords used together
with their identification numbers is given in Table 3-2. Articles that were
not relevant to the study were discarded.

050
Newton, E. H ■_________________________ ___________________________________
"Survey of Condenser Tube Life in Salt Water Service"
The International Nickel Power Conference, Wrightsville Beach,
N. C. (May 1968)

1.2, 4.1, 5.0, 8.1

A survey of the service life of condenser tube alloys in salt water service
was made during 1966 by Arthur D. Little, Inc. for the Office of Saline
Water, U.S. Dept, of the Interior. It is part of a continuing study being
carried out for OSW of the corrosion behavior of materials used in the con­
struction of large multistage flash distillation type desalination plants.
Electric power plants located on or near the sea coasts of the U.S. pro­
bably pump more sea water (or brackish water) collectively through their
turbine condensers than any other single or group of heat exchanger users.
Therefore, most of the effort in this survey was directed toward collecting,
assembling and interpreting data generously given us by a significant por­
tion of the electric utility companies who operate plants using sea or
brackish water for cooling.

Figure 3-1. Typical Catalog File Card

(original is 5 by 8 inches)

As each article was processed, its list of references was reviewed against the
existing catalog file. This produced an additional 100 references.

After nearly all of the known articles had been cataloged, etc. the informa­
tion, except the abstract, was transferred from the file cards to punched cards
to enable more efficient sorting, selection, and listings of the accumulated
bibliography.

3-3
 Table 3-2

LITERATURE SEARCH
KEYWORD LIST

1 CIRCULATING WATER TYPES 5 MATERIALS

1.1 Freshwater 5.1 Aluminum

1.2 Seawater 5.2 Admiralty
1.3 Brackish Water 5.3 Aluminum Brass
1.4 Polluted Seawater 5.4 Aluminum Bronze
1.5 Polluted Brackish Water 5.5 Brass
1.6 Cooling Tower 5.6 Copper and Its Alloys
1.7 Cooling Pond/Spray Canal 5.7 Copper-Nickel
1.8 Cooling Water Inlet 5.8 Stainless Steel
5.9 Titanium
CONDENSER COOLING SIDE TREATMENT 5.10 Analysis

2.1 Fouling 6 FABRICATION METHODS

2.2 Chlorine
2.3 Corrosion Inhibitors 6.1 Tube Rolling
2.4 On-Line Mechanical Cleaning 6.2 Tube Welding/Explosive Welding
2.5 Off-Line Mechanical Cleaning 6.3 Tube Repair
2.6 Off-Line Chemical Cleaning 6.4 Coatings
2.7 Inlet Inserts
7 LEAK DETECTION/LOCATION
CONDENSER DESIGN
7.1 Foam Method
3.1 Heat Transfer 7.2 Eddy Current/Probalog
3.2 Efficiency 7.3 Condensate Conductivity
3.3 Fouling Factor
3.4 Corrugated Tubes 8 MATERIAL TESTING
3.5 Thin Wall Tubes
3.6 Auxiliary Equipment 8.1 Generating Stations
3.7 Steam Dumping 8.2 Test Laboratories

CONDENSER FAILURE MODES 9 CONDENSER FAILURE IMPACT

4.1 Corrosion 9.1 Economic (Cost of Retubing)

4.2 Erosion 9.2 Physical
4.3 Missiles/Mechanical 9.3 Failure Rate Data
4.4 Vibration 9.4 Troubleshooting
4.5 Misoperation
4.6 Pressures 10 ENVIRONMENTAL IMPACT
4.7 Joints
4.8 Ammonia Corrosion 10.1 Bioassays
4.9 Failure Analysis 10.2 Pollution

3-4
A second punch card file was established for unpublished literature such as
internal memos and reports, which were obtained from the many contacts made
with representatives of utility companies, condenser manufacturers, tubing
manufacturers, etc.

A third punch card file was established for manufacturers' literature and
catalog information related to condensers, condenser tubing, cleaning devices,
and other accessory or peripheral equipment.

Results

The final bibliography, which contains 391 references listed numerically by

accession number, is given in Appendix A. In this report, literature
references are cited by this number.

The same bibliography listed alphabetically by author is given in Appendix B.

A keyword listing is given in Appendix C. In this listing, under each keyword

descriptor number (Table 3-2), there is given the accession number of all
reference articles to which the keyword is applicable. By successive cross­
comparison of the accession number listings under two or more keyword headings,
articles of increasingly specific subject content can be isolated.

Copies of all articles in the bibliography (except those with an asterisk

following the accession number) have been transferred to EPRI, together with
the catalog files and the punch card deck.

While all pertinent articles published since 1960 are not included in the
bibliography, the search techniques were sufficiently comprehensive to have
discovered a high proportion of the sought articles. In compiling the biblio­
graphy, only those articles having practical application to condenser tubing
failures, condenser cleaning, and similar subjects related to the immediate
scope of the study were included. Articles related to areas such as heat
transfer, the more theoretical aspects of the metallurgy of condenser tube
alloys, etc, were excluded.

3-5
The survey of the open literature revealed no comprehensive report of work
comparable to the present study. The closest analog is the interesting and
valuable work of Newton and Birkett (Ref 46 and 50) and Newton, Birkett, and
Ketteringham (Ref 350). These surveys were sponsored by the Office of Saline
Water, U.S. Department of Interior and were directed towards finding suitable
tubing materials for seawater desalination plants. Most of the data in the
first report (Ref 46) were drawn from the experience of power stations located
at coastal sites. Accordingly, the study was limited to saline cooling waters
and to alloys in common use in 1967.

MATRIX FOR SELECTION OF FIELD SURVEY STATIONS

Introduction

A crucial concern in making any survey based on a selected sample is that the
surveyed sample must be representative of the whole if the results and con­
clusions derived from the sample are to have broad application.

Even a cursory review of the general parameters related to condenser tube

deterioration suggests six or eight of prime significance (e.g., tube material,
circulating water composition, etc). Subdivisions and combinations of these
pointed to the need for a large survey sample. At the same time, however,
limitations on schedule and budget placed a sharp constraint on the number of
stations that could be included in the survey. Accordingly, a selection pro­
cess had to be devised that would produce a representative sample of the
significant parameters in sufficient quantity to form an adequate data base
for analyses yet would be appropriate to the available time and money. To
accomplish this critical task, a station selection matrix was prepared.

Station Selection Matrix

The station selection matrix grid contained 87 information fields of design,

operation, and historical condenser performance data for each of 264 generat­
ing units at 87 power stations.

Many factors were considered in the preparation of the station selection matrix.
Two important goals were: 1) to obtain a broad survey of operating experience
with large, modern utility condensers; and 2) to emphasize in the survey new
systems and materials that could be expected to be used in future applications.

3-6
EPRI in their Request for Proposal restricted the field survey to nuclear units
larger than 100 MWe, fossil units larger than 600 MWe, and plants containing a
significant number of titanium condenser tubes. However, because most of these
plants are less than 5 years old and nearly all are less than 10 years old, it
was decided to include a number of smaller fossil units. This was expected to
reveal the performance characteristics of units whose ages exceeded 15 years.
A detailed list of information items pertinent to the selection of stations was
prepared. Originally it was intended that the matrix would be completed solely
from data available in the literature, governmental reports, Bechtel files,
etc, but it became evident that additional information was required, especially
with regard to the availability and quality of station records and the stations'
condenser leakage history. To obtain this information, it was decided to con­
duct a mail survey of a representative number of stations.

MAIL SURVEY

Selection of Sample

In the United States there are about 1000 steam-electric generating stations.
However, 95 percent of the electrical output comes from about 600 stations
which have some 2000 generating units. It was decided that for the purposes
of this study, a representative profile would be obtained if 15 percent (about
90) of these stations were included in the mail survey.

Consistent with the EPRI criteria, a list was made of all stations with nuclear
units over 100 MWe (turbine nameplate rating), all fossil stations with one or
more units over 600 MWe, and all units with a significant portion of titanium
tubes in the main condenser. Only stations in commercial service prior to
January 1, 1975 were included. This list was supplemented with older, smaller
fossil stations chosen at random. The completed list contained 97 stations
with 290 generating units. Fifty-six operating utilities were represented in
the list.

The distribution of these stations with regard to parameters important to the

study will be described in the summary section below.

3-7
Mail Survey Form

To elicit a timely response from the maximum number of stations, a short,

two-page mail survey form was developed. This form placed emphasis on those
data which would be important in selecting stations for field survey but were
not readily obtainable otherwise. The mail survey form is attached as
Appendix D.

Mail Survey Response

The response to the mail survey questionnaire was excellent. In keeping with
the cooperative nature of the electric utility industry, a great majority of
the solicited stations responded promptly. A telephone follow-up campaign pro­
duced replies from most of the others. Eighty-seven stations with 254
generating units responded. This response was 89.7 percent of the stations
and 91.0 percent of the generating units to which the questionnaire had been
mailed.

The data obtained by the mail survey fulfilled the original objective and
enabled completion of the detailed Station Selection Matrix.

Summary

A partial Station Selection Matrix is included as Appendix E. All station

names, locations, and other identifying data have been deleted. Nonetheless,
the raw data base could be valuable to those who may wish to analyze it
for results beyond those given in this report. The descriptions of the
matrix information fields together with the abbreviations used in encoding
are given in Appendix F. The nine information fields marked with an
asterisk are not included in Appendix E.

A profile of the 264 generating units in the selection matrix with regard to
parameters relevant to the deterioration and failure of condenser tubes is
given in Tables 3-3 and 3-4. Table 3-3 shows the distribution by age for each
fuel type. Table 3-4 shows the distribution of condenser tube material in the
condensing section for each cooling water classification.

3-8
 Table 3-3

STATION SELECTION MATRIX

AGE DISTRIBUTION BY FUEL TYPE
Numbers of Generating Units

AGE RANGE FUEL TYPE PERCENT IN

YEARS BWR PWR FOSSIL ALL AGE RANGE
1-4.9 8 18 50 76 28.8
5.0- 9.9 4 5 37 46 17.4
10.0-14.9 0 2 34 36 13.6
15.0-19.9 1 0 32 33 12.5
20.0-29.9 0 0 56 56 21.3
>30 0 0 17 17 6.4
ALL AGES 13 25 226 264
PERCENT
EACH FUEL 4.9 9.5 85.6 100.0

Additional findings from the mail survey are reported in subsequent sections.

Table 3-4
STATION SELECTION MATRIX
CONDENSER TUBE MATERIAL BY COOLING WATER CATEGORY
Number of Generating Units

CONDENSER COOLING WATER CATEGORY*

** PERCENT
TUBE EACH
MATERIAL* FRESH CLOSED SALT ALL WATERS ALLOY

Admiralty*** 103 16 2 121 45.8

SS Type 304 25 22 0 47 17.8
SS Type 316 0 0 1 1 0.4
90-10 Cu-Ni 5 3 14 22 8.3
70-30 Cu-Ni 0 0 6 6 2.3
Aluminum Brass 0 0 37 37 14.0
Aluminum Bronze 0 0 8 8 3.0
Titanium 0 1 9 10 3.8
Other 11 1 0 12 4.6
ALL MATERIALS 144 43 77 264
PERCENT EACH
WATER 54.5 16.3 29.2 100.0

*Present material in condensing section.

**Fresh includes only once-through cooling mode.
Closed includes all cooling towers, closed ponds, and other recirculating
systems.
Salt includes seawater and brackish water.
***A11 types of admiralty are included.
"^"includes red brass (5) , muntz (2) , arsenical copper (4) , and brass (1) .

3-9
FIELD SURVEY

Criteria for Selection of Survey Stations

By an analytical review of the station selection matrix, 30 stations were chosen

for field survey. Evaluation of the selected sample showed that 30 stations
gave adequate opportunity for collection of data on all parameters of signi­
ficance to the study within the time and budget constraints.

The following criteria were used in selecting the survey stations and were
applied approximately in the order listed.

Fuel Type. Emphasis was placed on nuclear stations since their relative numbers
will probably increase in the future. Eleven nuclear stations were chosen.
Of these, four were BWRs and seven were PWRs, which is nearly the existing ratio
between the two major types of nuclear systems.

Cooling Water Category. Stations having closed cycle cooling systems were pre­
ferred for study because such systems can be expected to be predominant in the
future. Also, emphasis was given to stations that used brackish water or sea­
water for cooling because this environment has had a higher incidence of con­
denser tube failures and because the impact of condenser leakage on plant
operations is more severe.

Condenser Tube Material. Special care was taken to assure that all of the
principal condenser tube alloys were adequately represented in the survey
sample. However, the sample selection was weighted to emphasize stations that
had a significant number of titanium tubes as well as stations having stainless
steel or copper-nickel tubes in the main condensing section.

Unit Size and Ages. In selecting stations with fossil fuel units larger than
600 MWe, preference was given to those which had older, smaller units on the
site in order to broaden the age data base while keeping the number of survey
trips to a minimum. Also preferred were stations where two or more condenser
tube materials were or had been in service under the same conditions.

Records Quality and Availability. Stations were preferred if their mail survey
response indicated that they had acceptable records that would be available to
the survey team. The mail survey responses to the records-related questions
were rated and reduced to a letter grade. The results are plotted in Figure 3-2.

3-10
 FIELD SURVEY

MAIL SURVEY
FOSSIL UNITS
> 600 MWE

MAIL SURVEY
NUCLEAR UNITS

MAIL SURVEY MAIL SURVEY

ALL UNITS J UNITS> 15 YEARS OLD

BEST POOREST
CONDENSER RECORDS QUALITY

Figure 3-2. Mail Survey Results of Condenser-Related Records

Condenser Manufacturer. The tentative selection list defined by the above

criteria was reviewed to ascertain whether the several condenser manufac­
turers were proportionately represented.

Condenser Failure History. Newer stations that reported few if any condenser
tube failures in the mail survey were not included in the field survey.

Station Location. The selection list was reviewed to ensure that excessive
time and travel would not be necessary.

As the field survey work progressed, the list of stations and survey schedule
was revised as dictated by early results and the availability of stations to
suit the survey schedule.

3-11
Characteristics and Validity of Survey Sample

The 30 stations surveyed had 76 generating units. The median age of the
generating units was 12.1 years, and the median generating capacity was
476 MWe. The 76 unit condensers included 116 condenser shells which contained
294 tubing bundles. In all, these condensers had 2,288,722 tubes with a total
surface area of 23,189,000 square feet. There were over 15,000 lineal miles
of condenser tubing.

Tables 3-5 and 3-6 give the same correlations for the Field Survey Stations as
were given in Tables 3-3 and 3-4 for the stations in the selection matrix.
Table 3-5 shows the number of generating units in each age range for each fuel
type. Table 3-6 shows the number of generating units using each condenser
tube alloy versus the cooling water category. Additionally, Table 3-7 shows
the unit age by generator size distribution for the field survey.

Table 3-5

FIELD SURVEY STATIONS

AGE DISTRIBUTION BY FUEL TYPE
Number of Generating Units

PERCENT IN
AGE RANGE FUEL TYPE EACH
YEARS BWR PWR FOSSIL ALL AGE RANGE

1-4.9 1 6 11 18 23.7
5.0- 9.9 4 3 15 22 29.0
10.0-14.9 0 0 14 14 18.4
15.0-19.9 1 0 14 15 19.7
20.0-29.9 0 0 7 7 9.2
> 30 0 0 0 0 0.0
ALL AGES 6 9 61 76
PERCENT
EACH FUEL 7.9 11.8 80.3 100.0

Notes
1. 36.7 percent of stations surveyed were nuclear.
2. 20 of the 61 fossil units surveyed were larger than 600 MWe.
3. 10 of the 15 nuclear units surveyed were larger than 600 MWe.

3-12
 Table 3-6

FIELD SURVEY STATIONS

CONDENSER TUBE MATERIAL BY COOLING WATER CATEGORY
Numbers of Generating Units

CONDENSER COOLING WATER CATEGORY*

** PERCENT
TUBE EACH
MATERIAL* FRESH CLOSED SALT ALL WATERS ALLOY

Admiralty*** 29 6 0 35 46.1
SS Type 304 7 3 0 10 13.2
SS Type 316 0 0 1 1 1.3
90-10 Cu-Ni 2 0 11 13 17.1
70-30 Cu-Ni 0 1 1 2 2.6
Aluminum Brass 0 0 1 1 1.3
Aluminum Bronze 0 0 5 5 6.6
Titanium 0 0 9 9 11.8
Other 0 0 0 0 0.0
ALL MATERIALS 38 10 28 76
PERCENT
EACH WATER 50.0 13.2 36.8 100.0

*Present material in condensing section.

**Fresh includes only once-through cooling mode.
Closed includes all concentrating cycles with freshwater makeup.
Salt includes all seawater and brackish water cycles.
***A11 types of admiralty are included.

3-13
 Table 3-7

FIELD SURVEY STATIONS

AGE DISTRIBUTION BY GENERATOR SIZE
Number of Generating Units

PERCENT
AGE RANGE GENERATOR NAMEPLATE RATING, MWe EACH
YEARS <49.9 50-99 100-299 300-599 600-999 >1000 AGE RANGE

1- 4.9 0 0 0 5 11 2 23.7
5.0- 9.9 0 0 1 8 12 1 29. 0
10.0-14.9 0 0 5 5 4 0 18.4
15.0-19.9 1 1 11 2 0 0 19.7
20.0-29.9 3 2 2 0 0 0 9.2
>30 0 0 0 0 0 0 0.0
ALL AGES 4 3 19 20 27* 3**
PERCENT
EACH SIZE 5.3 3.95 25.0 26.3 35.5 3.95 100.0

*Includes five BWR and five PWR units.

**A11 three units are fossil fueled.

The station selection matrix does not present a typical profile of all genera­
ting units in the United States. It is, however, representative of large
modern power generating units and provides an excellent yardstick by which to
assess the validity of the selected field survey sample. Comparisons of the
generating units in the station selection matrix with those in the field sur­
vey are made in Tables 3-8, 3-9, and 3-10.

Table 3-8 shows the percentage of units in each data base for the correlation
between unit age and fuel type. As would be expected from the criteria used
for their selection, the field survey stations show a higher ratio of nuclear
to fossil units and an emphasis on stations in the 5-to-20-year age range.
The comparison of condenser tube material with cooling water category given
in Table 3-9 shows the heavier weighting given to 90-10 Cu-Ni and titanium.
It also shows that less than the intended weighting was given to stainless
steel alloys, especially in closed cooling systems. This was largely because
such stations are relatively new and have insufficient meaningful experience.

3-14
However, this was compensated in part by the many stainless steel air removal
sections that were surveyed. As will be brought out in Section 5, more infor­
mation was obtained for the aluminum brass/bronze alloys than is indicated on
the list of tube materials. At a number of the surveyed stations, data were
obtained on original tube sets of these alloys which have been replaced with
another material. Admiralty, which is by far the predominant material used
in freshwater condensers, was included in the field survey in the same ratio
as it exists in the selection matrix. The field survey emphasized seawater
cooled systems as intended but contained a lower ratio of closed cooling
systems because most such stations have been built more recently.

Table 3-10 tabulates the distribution of condenser manufacturers for the mail
survey and the field survey. The distribution in the field survey was
representative for the major manufacturers.

Table 3-8

COMPARISON OF STATION SELECTION MATRIX WITH FIELD SURVEY

AGE DISTRIBUTION BY FUEL TYPE
Percent of Generating Units in Each Sample

FUEL TYPE
AGE RANGE BWR PWR FOSSIL ALL
YEARS SSM* FIELD** SSM FIELD SSM FIELD SSM FIELD

1-4.9 3.0 1.3 6.8 7.9 19.0 14.5 28.8 23.7

5.0- 9.9 1.5 5.3 1.9 3.9 14.0 19.8 17.4 29.0
s

0.0 0.0 18.4

o
o

&

0.8 0.0 12.9 13.6 18.4

15.0-19.9 0.4 1.3 0.0 0.0 12.1 18.4 12.5 19.7
20.0-29.9 0.0 0.0 0.0 0.0 21.2 9.2 21.3 9.2
30 0.0 0.0 0.0 0.0 6.4 0.0 6.4 0.0

ALL AGES 4.9 7.9 9.5 11.8 85.6 80. 3 100.0 100.0

*Station Selection Matrix contains 87 stations with a total of 264 generating

units.
**Field survey includes 30 stations with a total of 76 generating units.

3-15
 Table 3-9

COMPARISON OF STATION SELECTION MATRIX WITH FIELD SURVEY

CONDENSER TUBE MATERIAL BY COOLING WATER CATEGORY
Percent of Generating Units in Each Sample

CONDENSER COOLING WATER CATEGORY

TUBE FRESH CLOSED SALT ALL WATERS
MATERIAL ** FIELD*** SSM
SSM* FIELD SSM FIELD SSM FIELD

Admiralty! 39.0 38.2 6.0 7.9 0.8 0 45.8 46.1

SS Type 304 9.5 9.2 8.3 4.0 0 0 17.8 13.2
SS Type 316 0 0 0 0 0.4 1.3 0.4 1.3
90-10 Cu-Ni 1.9 2.6 1.1 0 5.3 14.5 8.3 17.1
70-30 Cu-Ni 0 0 0.4 1.3 1.9 1.3 2.3 2.6
Aluminum
Brass 0 0 1.5 0 12.5 1.3 14.0 1.3
Aluminum
Bronze 0 0 0 0 3.0 6.6 3.0 6.6
Titanium 0 0 0.4 0 3.4 11.8 3.8 11.8
Other 4.2 0 0.4 0 0 0 4.6 0.0

ALL
MATERIALS 54.6 50.0 16.2 13.2 29.2 36.8 100.0 100. 0

*Present material in condensing section.

**Station Selection Matrix contains 87 stations with a total of 264 generating
units.
***Field survey includes 30 stations with a total of 76 generating units,
tlncludes all types of admiralty.

3-16
 Table 3-10

COMPARISON OF STATION SELECTION MATRIX WITH FIELD SURVEY

DISTRIBUTION BY CONDENSER MANUFACTURER
Percent of Generating Units in Each Sample

CONDENSER STATION SELECTION FIELD

MANUFACTURER MATRIX* SURVEY**

Allis-Chalmers 0.4 5.3

Baldwin-Lima-Hamilton 3.4 3.9
CH Wheeler 4.9 7.9
Foster-Wheeler 22.3 25.0
Ingersoll-Rand 21.6 22.4
Lummus 0.8 1.3
Maryland Drydock & SB 0.4 0.0
Southwest Engineering 2.7 2.6
Westinghouse 16. 7 26.3
Worthington 8.3 5.3
Yuba 1.1 0.0
Not Indicated 17.4 0.0

NUMBER OF UNITS 264 76

^Station Selection Matrix contains 87 stations with a total

of 264 generating units.
**Field survey includes 30 stations with a total of 76 genera­
ting units.

3-17
Field Survey Form

A comprehensive and orderly field survey form was developed for use by the
survey teams. The form was designed with the following purposes in mind:

• To serve as a checklist and guideline for the efficient

conduct of each survey
• To provide for the collection of all pertinent data and
thereby minimize the need for followup work
• To avoid the collection of peripheral or irrelevant data
• To promote standardization of data collection by different
survey teams.

The 30-page field survey form was organized in seven sections:

1. General Station Data

2. Document Request List
3. Condenser Design Data
4. Chemistry - Cooling Water and Turbine Cycle
5. Condenser Operation, Maintenance, and Repair
6. Condenser Leakage Impact
7. Additional System and Equipment Data

A form was completed for each generating unit surveyed at the station. The
entire form is given in Appendix G.

Procedure for the Field Surveys

The field surveys were conducted by multidisciplined teams of 1 to 3 engineers

Metallurgists, mechanical engineers, and water chemistry specialists were
assigned to each team in accordance with the conditions anticipated at each
station as determined from the station selection matrix and preliminary
telephone contacts. In some instances, one senior staff specialist of the
most appropriate discipline conducted the entire survey.

Each selected station was telephoned to arrange the schedule and advised of
the information needed. When time allowed, the survey form and other infor­
mation were sent to the station in advance to assist their preparation for
the survey. All of the survey stations were very cooperative. The typical
visit consisted of two engineers on site for two days.

3-18
After the first several field surveys, it became evident that detailed quanti­
tative data for all of the desired areas could not be obtained in the 3 to 5
days devoted to each field survey. It was found that nearly all stations had
quite accurate records of the number (and usually the location) of plugged
tubes. Some stations did not log the date on which the tubes were plugged.
In other cases, the only time-oriented record consisted of colored symbols on
tubesheet drawings where the color indicated the date. Such drawings cannot
be readily reproduced and to make a take-off tabulation would require excessive
time. It was also found that reliable, detailed records of the cause of the
tube failure were not available at many stations.

It was also found that only a few stations had calculated and recorded the
inleakage flow rate for individual condenser leaks. At most stations, the
data to enable such calculations probably are in the stored records. However,
to bring together (from various depositories) the several necessary records,
and to time-synchronize them would require approximately one week of effort
for each condenser-year studied.

In conjunction with most surveys, a visit was made to the headquarters

engineering department of the operating utility. This enabled collection of
valuable data not attainable on site with regard to not only the surveyed
station but also other stations in the system. A wealth of information was
obtained in the form of operating records, company reports, literature
references, etc.

To supplement the information gained from the operating utilities, several

meetings were held with manufacturers of condensers and condenser tubing.
Again, excellent cooperation was accorded to the field survey personnel and
an abundance of data, research reports, literature references, and design
experience were obtained.

DATA MANAGEMENT SYSTEM

To cope with the literally hundreds of thousands of information items obtained

in the course of the study required development of an effective computerized
data management system. Such a system had to enable efficient and rapid sort­
ing, arraying, correlating, averaging, tabulating, etc . , of the data bank. To
do this, a matrix tabulation was devised to contain all of the relevant alpha­
numeric data from the field survey which were subject to encoding.

3-19
The field survey matrix is schematically illustrated in Figure 3-3. The
matrix column headings consist of 198 information fields. Each of these is
defined and the abbreviations used of encoding are given in the Description of
Information Fields in Field Survey Matrix, Appendix H. The information fields
were ordered roughly by subject areas. From left to right these are: Station
Data; Unit Data; Condenser Design Data; Condenser Tubing Data; Failure Rate Data;
Leak Detection, Location, and Repair Data; and Data on other Equipment.

The entered data on each line in the matrix are descriptive of an individual
tubeset. As discussed in Section 2, a condenser has three primary service
zones or regions, namely: impingement, air removal, and condensing. In many
condensers, the tubing material used in the impingement and/or air removal
zone is different from that used in the main condensing zone. The tube failure
mechanism and the rate of tube failures can also be markedly different in each
zone as well as for each material. Furthermore, a number of the condensers
surveyed had been retubed in whole or in part one or more times. It follows
then that to have rational and comparable tube failure data, the tubes in any
given condenser must be broken down into individual tubesets of the same
material and in the same service.

For the condensers surveyed in the field, 684 tubesets were identified; hence
the final matrix contained 684 lines. The total matrix had space available for
135,432 information items. Many of these were repetitive for all tubesets at
a given station or for a given generating unit, and for many others the
information was not available.

The raw data from the field survey forms were transferred to punch cards.
Twelve punch cards were required to accommodate the 198 information fields.
A set of 12 cards was prepared for each of the 684 tubesets. The deck of
8208 punch cards was used to load the matrix into the computer.

The data bank was manipulated for sorts, arrays, and various operations by
System 2000* (version 2.8), a broadly applicable data management system
developed by MRI Systems Corporation of Austin, Texas. System 2000 was

*System 2000 is a registered service mark.

3-20
 ------ lOlii
198 INFORMATION FIELDS (12 PUNCH CARDS)
(960 COLUMNS)
•4---------------------------------- ►>] 1 SUBJECT AREA
(1 TO 320 COLUMNS)
•«------------------H 1 PUNCH CARD (80 COLUMNS)

-*—►) 1 INFORMATION FIELD (1 TO 19 COLUMNS)

h«— 1 COLUMN
1
i. J
1 UNIT 1 TUBESET T
(1 TO 24) (1 LINE) 1
(TUBESETS) (1 PUNCH CARD)
1r

1 STATION
(1 TO 5 UNITS)

ir

TOTAL MATRIX
CONTAINS:
30 STATIONS
(76 UNITS) 135,432 INFORMATION
(684 TUBESETS) ITEMS
8,208 PUNCH CARDS
656,640 PUNCH SLOTS

U
A/

Figure 3-3. Data Management Matrix

3-21
chosen because of its suitability for the problems at hand and because it is
available through many time-sharing or data management firms. The punch card
deck together with a magnetic tape containing the data bank and the adoption
routines for System 2000 have been turned over to EPRI.

A partial field survey matrix is given in Appendix I. All station

names, locations, unit numbers, and other readily identifiable items
have been deleted; thirty-one information fields have been deleted.
These are designated by an asterisk preceding the field number in the
listing given in Appendix H. To make this matrix more useful to the
reader, it has been sorted sequentially by tubing material (field 49),
condenser service area (field 170), cooling water category (field 171),
and arrayed in ascending numerical order of the terminal, total failure
rate (FRT - field 124). The terminology is defined in Appendix H and
discussed in Section 5.

The station selection matrix was prepared by essentially the same technique
and also adapted for manipulation by System 2000. A punch card deck and
magnetic tape similar to that for the field survey matrix also has been
delivered to EPRI. This facilitates future use of either data base by EPRI
or its designee.

3-22
 Section 4

CONDENSER OPERATION AND MAINTENANCE FINDINGS

INTRODUCTION

A condenser may affect plant performance as a result of the following

deficiencies:

• Reduction of heat transfer capability

• Reduction of deaeration efficiency
• Loss of integrity (inleakage of air or water)

To protect against potential problems, the condenser and its related systems
are operated and maintained in accordance with the prevailing conditions.
Specific plant parameters such as circulating water composition and condensate
water quality specifications affect local operation and maintenance practices.
More generally, however, such practices reflect broader needs and considera­
tions which are reviewed below.

The major study findings of interest in condenser operation and maintenance are
reviewed in this section of the report. The following topics are addressed:

• Circulating water inleakage and air inleakage from the view­

point of their impact on the plant; their frequency, size,
and duration; the methods used for their detection, location,
and repair; and the procedures that seem most effective for
their control.
• Various operating and maintenance considerations found to be
significant in ensuring condenser integrity and high perfor­
mance, such as tube cleaning methods used, including their
required frequency of application and their effectiveness.
• An evaluation of current plant record-keeping practices.

CIRCULATING WATER INLEAKAGE

Water Leakage Impact

The mail and field surveys confirmed the general background of information
collected as part of the literature search to the effect that circulating

4-1
water inleakage is detrimental to power plants for the following reasons:

• Circulating water impurities that enter the steam cycle as

the result of condenser leakage contaminate the condensate,
feedwater, steam generator water, and steam. Above some
given leakage rate, which is a function of the impurity con­
tent of the circulating water and of the cycle water quality
specifications, the faulty condenser or condenser section
must be isolated and normal operation cannot be resumed until
the leakage is repaired.
• The plant must be operated at part load or even completely
shut down for some period of time. The leakage impact is
therefore the cost of the repair plus the costs resulting from
loss of plant availability and loss of energy generation.
• During plant operation with circulating water inleakage,
additional costs are incurred to maintain the steam cycle
streams within acceptable impurity limits. Cycle water
quality is maintained by increasing the steam generator blow­
down, by operating condensate polishers, or by a combination
of the two. However, steam generator blowdown wastes valu­
able energy and requires some form of chemical processing,
either for recycle or for replacement makeup water and, in
some cases, for treatment prior to off-site discharge.
Condensate polishers require either chemical regeneration or
replacement of their ion exchange resins. Whatever purifica­
tion process is used, additional operating costs are incurred
until the leak is repaired.
• Finally, a condenser may deteriorate in service to the point
that either its maintenance requirements make it uneconomical
to operate as is, or its performance falls below acceptable
levels in some regards (e.g., its air removal section may
have so many plugged tubes that it does not collect and con­
centrate the noncondensible gases as effectively). At this
point, the condenser is usually retubed totally or in part.
As illustrated in Figures 4-1 and 4-2, however, retubing is
a major operation which is costly in terms of both direct
costs and loss of plant output.

In addition to the above impacts, condenser water inleakage may lower the per­
formance and may accelerate the general degradation of other steam cycle
components, mostly as the result of fouling and/or corrosion. The field
survey findings strongly suggest that, with regard to sensitivity to leakage
degradation, power plants should be grouped according to fuel type.

• Fossil fueled plants showed little apparent degradation. Some

of the plants surveyed indicated that circulating water
inleakage had contributed to some extent to solids deposition
in the boiler. A few other plants estimated that part of the
turbine deposits could be traced back to the same cause. It
is very difficult to correlate water leakage to corrosion or
fouling phenomena; yet it appears from the survey results

4-2
 ssH

Figure 4-1. Condenser Retubing Work Area

Figure 4-2. Condenser Retubing — Tube Pulling Operation

4-3
 7

that fossil fueled plants that have been operated generally

within their water quality specifications show no evidence
circulating water inleakage impact.

• The survey findings suggest that the above conclusion is also

applicable to Boiling Water Reactor plants (BWRs) however,
the data collected and analyzed for BWR plants was not
sufficiently broad and detailed to allow a conclusion to be
reached. Also, one of the BWR plants surveyed had major
corrosion damage after an extended chemistry excursion during
a period of leakage.

• The survey found that some Pressurized Water Reactor plants

(PWRs) had significant damage in the steam generators, per­
haps due in part to continued operation with condenser water
inleakage. This observation was not the result of an evalua­
tion by the survey team but was based on statements by plant
personnel.

Water Leakage Frequency, Size, and Duration

The frequency of condenser water leakage occurrences is an indication of con­

denser integrity status, and it varies over a very wide range. Some of the
plants surveyed had leakage occurrence frequencies in the range of once a
decade; others had frequencies in the range of once a week. However, leakage
frequency is not always a significant parameter for the following reasons:

• Many plants can continue to operate at full power, well with­

in water quality specifications, even when the condenser is
leaking, if they have relatively clean circulating water and
have an efficient and high-capacity cycle purification system
(steam generator blowdown and/or condensate polisher).

• Some plants can use sawdust to plug leaks on-line that would
otherwise have to be repaired promptly.

• In most plants it is desirable during maintenance operations

to plug not only the leaking tubes but also to guard plug the
tubes that have a relatively high probability of becoming
leakers (based on such considerations as proximity to
defective areas or eddy current results).

Thus, leakage occurrence frequency is not a simple and direct indicator of

plant outage due to condenser inleakage. It is also a difficult parameter to
measure, and it is not generally tracked by operating plants. It could not
be evaluated with accuracy from the data collected by the surveys.

The size or flow rate of condenser inleakage that can be tolerated during
plant power operation also varies over a very wide range, mostly as a function
of the circulating water composition and of the plant cycle purification
system's capacity and efficiency. For example, one of the fossil plants

4-4
surveyed had been operating for several weeks with a completely sheared tube
that was leaking more than 200 gallons a minute; on the other hand, one of
the nuclear plants surveyed was ramped down for repair after it had operated
for a few hours with a leak of less than 0.01 gallon a minute. Thus, no pre­
cise statement can be made with regard to acceptable leak rates unless all the
relevant plant parameters are known.

It was observed that plants cooled by seawater or brackish water are generally
limited to leakage rates in the range of a few tenths of a gallon a minute or
less, but open-cycle freshwater plants can continue full power operation with
leakage rates of several tens of gallons a minute. Nuclear plants generally
tolerate less leakage than fossil plants.

Condenser water leakage durations are also distributed over a wide range and
are a function of the parameters referred to above. The data collected during
the surveys showed that the rate of growth of any given leak and therefore its
allowable duration was a highly variable parameter even within a given plant.
In addition, the data collected suggest that leaks grow very erratically and
that with circulating water of high suspended solid content, temporary spon­
taneous leak plugging is not uncommon. Despite the large variability referred
to above, seawater- or brackish water-cooled plants generally tolerate leakages
lasting between a few hours and a couple of days, but open-cycle freshwater
plants can usually continue unrestricted but prudent operation with water
leakages lasting from several weeks to a few months.

Water Leakage Detection

Condensers are usually monitored for possible water inleakage by a combina­

tion of methods which fall into two categories. In the first category are
those methods used during power operation. They are based on monitoring
selected cycle streams for chemical changes. In the second category are those
methods used during plant shutdown. They are generally highly sensitive in
terms of detectable leak flowrate and are particularly valuable because they
allow the locations of the leaks to be identified. Accordingly, they are
described in the subsection, "Water Leakage Location," (see page 4-11) and the
following paragraphs address only the on-line detection methods.

On-line methods are based on detecting the changes in cycle water chemistry
that are induced by condenser leakage. When expressed in terms of leak flow
rates, the sensitivity of any given method is directly related to the

4-5
circulating water composition, the inherent sensitivity of the equipment
used, and to normal cycle chemistry fluctuations. From a practical view­
point, the dominant consideration in monitoring water leakage is the ability
to detect any leakage well before it has a significant effect on the plant
chemistry. In particular, the size of the leak is not significant except as
an indication of whether or not the leak can be readily located when the con­
denser is opened. There is little merit in developing and implementing a
leak detection technique that has a sensitivity much greater than that re­
quired to identify the smallest leak which can be located and repaired. It
is necessary to provide for the detection of the smallest leaks which can be
located and repaired. Thus, most plants use detection and location techniques
that have approximately the same sensitivities (see Figure 4-3).

SAMPLE SIZE: 108 UNITS

(From Mail and Field Survey)

WATER LEAK DETECTION TO LOCALIZATION RATIO

Figure 4-3. Plants Surveyed that Achieve a Given Condenser Water Leakage
Detection-to-Localization Rate

4-6
In the following discussion, leakage detection sensitivities are defined as
the changes in readings of monitored parameters that the plants considered to
be meaningful indicators of condenser leakage. The discussion is presented
in two parts; the first considers BWR plant cycles which use neutral water,
and the second considers fossil and PWR plants which maintain a high pH
throughout their cycle.

BWR Stations. All BWR plants use condensate straight conductivity monitors as
the major indicators of potential condenser water leakage. These instruments
have a low purchase and maintenance cost and are very sensitive and dependable.
They are always installed at the condensate pump discharge and, in most cases,
a number of cells are inserted in the hotwells. During plant steady power
operation, this method is capable of producing the following sensitivities:

• A conductivity increase of 0.02 to 0.04 umho/cm is generally

considered to be a reliable indication of abnormal condensate
impurity levels and to justify an on-line investigation of
possible contamination (including instrumentation checks,
comparison of instrument readings, check of makeup water
quality, and analyses of selected samples).
• An increase of 0.05 to 0.10 pmho/cm is generally considered
to be a reliable indication of condenser water inleakage and
to justify a plant load reduction for location and repair at
the earliest convenient time.

In addition, BWR plants always monitor reactor water very closely, particularly
for chlorides. This gives an independent measurement of condenser water
inleakage.

Fossil Fueled and PWR Plants. These plants use a variety of leakage detec­
tion techniques which are briefly described below. The frequency of utili­
zation is also included.

Condensate Cation Conductivity. This is used as the primary condenser leakage

indicator at 55 percent of the plants surveyed and as a secondary indicator at
most of the other plants. It is based upon measuring the conductivity of the
effluent from a hydrogen form cation exchange resin. (See Figure 4-4.)
Because it includes such a resin column, it has a slightly higher purchase
cost and requires somewhat more attention and maintenance than the straight
conductivity method. However, cation conductivity is also more sensitive than
straight conductivity and is reported as dependable. As a result, it is a
very efficient monitoring technique in modern fossil fueled and PWR plants

4-7
 Figure 4-4. Cation Conductivity Instrument Resin Columns

which maintain a fairly high pH in their cycle and cannot use straight conduc­
tivity monitors. Consistent with BWR plant monitoring schemes, it is applied
at least at the condensate pump discharge and, in difficult cases, at a number
of sample lines taking samples directly out of the condenser. The survey
indicates that, during plant steady power operation, the sensitivities are as
follows:

• A cation conductivity increase of 0.05 to 0.10 ymho/cm is

generally considered to justify some on-line investigation of
possible contamination.
• An increase of 0.10 to 0.20 ymho/cm is generally considered
to be a reliable indication of condenser leakage.

At each of the above levels, actions similar to those described for BWR plants
are usually implemented in fossil and PWR plants.

In addition to condensate monitoring, fossil and PWR plants which use drum-
type steam generators usually also monitor drum water and thus obtain a second
set of measurements that can be used as indicators of condenser water leakage.
The condensate impurity level at which condenser leaks are identified is shown
in Figures 4-5 and 4-6.

4-8
 i k

50
Sample Size: 15 Units
46%
(From Mail and Field Survey)

27%
PERCENT

20%

7%

0.05 0.2 0.8 3

CONDENSATE IMPURITY CONTENT, PPM

Figure 4-5. PWR Plants Surveyed That Detect Condenser Water Leakage
at a Given Bulk Condensate Impurity Content

4-9
 SAMPLE SIZE: 80 UNITS

(From Mail and Field Survey)

CONDENSATE IMPURITY CONTENT, PPM

Figure 4-6. Fossil Plants Surveyed (> ISOMWe) That Detect Condenser
Water Leakage at a Given Bulk Condensate Impurity Content

4-10
Straight Conductivity. This is used at 17 percent of the plants surveyed.
Older and smaller fossil plants which maintain their condensate at a pH only
slightly above neutral rely upon this method. In this application, the straight
conductivity method is used as described for BWR plants.

Other Methods. The remaining fossil and PWR plants surveyed (28 percent of the
total) use the following primary indicators for condenser water leakage monitor­
ing: silica (11 percent); sodium (6 percent); steam generator drum general
chemistry parameters (5 percent); steam generator drum chloride (3 percent);
hardness (2 percent); some combinations of the cycle chemistry parameters
(3 percent). These methods were selected for the most part because of rather
unusual conditions prevailing at the plants where they are used. As a rule
they are not of general interest and are not further discussed in this report.

The sodium method is not subject to the same limitation and is further
evaluated. Its sensitivity during steady power operation was reported to
be in the range of 1 ppb for a reliable indication of abnormal condensate
impurity level and 2 to 3 ppb for a reliable indication of condenser leak­
age. At many plants, the circulating water composition is such that the
sodium method would be capable of detecting and tracking leakages smaller
than those which could be monitored by cation conductivity measurements.

Until now the sodium method has not been applied to a large extent because it
has relied on equipment that has required extensive operating and/or mainte­
nance care (early sodium selective electrode instruments and atomic absorption
analyzers). Reportedly, both types of sodium instrumentation have recently
been substantially improved. In the future, condenser water leakage may be
monitored more frequently by measurements of selective cations, particularly
sodium.

Water Leakage Location

Two categories of methods are available for locating condenser water leaks.
Prelocation methods are applicable chiefly to large, modern condensers. They
are aimed at defining the general area of the leakage. Localization methods
are applicable to all condensers. They are aimed at defining the leak location
to the full extent required for leak repair. Some of these methods are also
used for detecting small leaks. All of them are normally implemented just
before the leak repair work is performed and as part of the same maintenance
operation.

4-11
 Prelocation and localization methods are based upon creating some pressure
differential between the steam side and the circulating water side and de­
tecting the resultant flow of water, air, or some other gas. The resultant
flow is directly related to the operating circulating water leakage, but it
is not equal to it. Many location methods rely on detecting the air inleak­
age induced by the condenser vacuum. Figure 4-7 shows this air inleakage
rate as a function of postulated leak hole size; it also shows, as a function
of hole size, the operating water leak rates for three typical water-box
pressures. The figure illustrates that the sensitivity of vacuum-based
location techniques is a function of operating water-box pressure when
expressed in terms of detectable operating water leak rate.

1.4-n

1.2 — — 3.0

FOR CIRCULATING
WATER PRESSURE OF 57 psia

1.0 - — 2.5
IN L E A K A G E R A TE , gpt

s<
2.0

.KAG E R A T E ,
0.8 — —

„ FOR CIRC WATER

PRESSURE OF 7.5 psia

0.6 - — 1.5 2

q:
LU <
<
£ FOR ATMOSPHERIC
0.4- — 1.0
CIRC WATER PRESSURE

0.2 — -0.5

EQUIVALENT HOLE DIAMETER, inch

Figure 4-7. Air and Water Inleakage Rates for a Condenser with
1.7 inch Hg Backpressure

4-12
Data collected during the surveys indicate that the following prelocation
techniques are most commonly used:

• In condensers that have several parallel circulating water

streams, the first prelocation step is to identify the stream
from which the leakage takes place. The plant load is
reduced to permit operation with one stream shut down and
each stream is temporarily isolated and drained. If the
purpose is to prelocate gross leakage, it is sufficient to
follow the leakage detection instrumentation response. When
the leak or leaks are uncovered, circulating water impurities
can no longer enter the cycle but air is allowed to enter the
condenser. The off-gas flowmeter, if available, responds
almost immediately to the change; however, it is not a very
sensitive instrument and can detect only very large leaks.
Water chemistry instrumentation is more sensitive, but its
response may be distorted for a time by a temporary buildup
of impurities in the condenser hotwell. The buildup takes
place between the time the circulating water flow is inter­
rupted and the time the leak is uncovered.
• As an alternative to the above, the faulty circulating water
stream may be identified by injecting freon into the desired
circulating water lines and setting a freon detector at the
condenser off-gas outlet. This method was reported by
several plants as being both dependable and sensitive enough
for their needs.
® In a refinement of the above two methods, plants that can
adjust and measure water box levels have used these measure­
ments to estimate the approximate height of the leaks. Neither
method is very effective for small or multiple leaks.
• Finally, some condensers are internally partitioned and
fitted with sample lines. Normal sample monitoring allows
the detection of leaking impurities before they are diluted
by the bulk of the condensate flow and allows identification of
the faulty condenser section. This identification must be
confirmed before undertaking any corrective action since
experience indicates that cross-contamination is not unusual
when a leak develops at or near a condenser section interface.

The surveys also show that most plants determine the exact location of condenser
water leaks by one or more of the following well-known localization techniques,
all of which require direct access to the condenser tubesheets:

• Vacuum-related methods are used with the condenser steam side

under normal operating vacuum. The most commonly used methods
consist of applying, on corresponding areas of opposite tube-
sheets, a material such as plastic film, detergent foam, wet
paper, or light paste which can seal tightly on the tubesheet
and be drawn in by the vacuum resulting from a leak. The
most successful are the use of plastic films or heavy foams.
Heavy foam is capable of locating water leaks as low as
0.01 gpm in condensers operating with low circulating water
pressure.

4-13
 • Other vacuum-related techniques of interest rely on pressure
gauges, air bubbles, flame indicators, or ultrasonic detec­
tors. The ultrasonic detectors are particularly effective in
locating large leaks when they are at or close to the tube-
sheet being surveyed. The method consists of scanning the
suspected tubesheet area with an ultrasonic sensing atmos­
pheric probe which has a peak sensitivity between 35 and
45 kHz. The probe detects the ultrasonic waves generated by
the leaking gas turbulence. Its operating frequency range
is that which experience has shown to result in the highest
signal-to-background ratio around most industrial facilities.
The sensitivity could not be accurately defined from the sur­
vey data. It appears to be in the gpm range for leaks located
in the tube central sections and to be better than one tenth
of a gpm for leaks located at or near the tubesheet (again
considering condensers operated with low circulating water
pressure).
• Hydrotest methods are used during complete plant shutdown by
filling the condenser steam side with water. This is some­
times done with dyed water which allows for a precise identi­
fication of all significant leaks. This is often used as
part of the normal condenser preventive maintenance program.
The condenser must be capable of accepting the relatively
large weight of water that is required.

Water Leakage Repair

Condenser water leaks almost always develop in or around the condenser tubes.
On rare occasions, leakage can occur through the tubesheet to shell joint.

Tube leaks are usually repaired by inserting a plug in the opposite ends of
the defective tube. Typical plugs are shown at the lower right in Figure 4-8.
The plugging operation is straighforward and usually presents no problems.

However, condensers operated on highly corrosive waters should be repaired

using only a plug material that has proved capable of withstanding the
circulating water environment.

Another well-known tube leak repair method is the addition of sawdust or a

cellulose-asbestos mixture. Its benefits are temporary and somewhat unpre­
dictable. It was extensively used until environmental protection regulations
banned its use for open-cycle circulating water systems. It is still in use
in at least one cooling tower plant. Its possible effect on the towers could
not be evaluated adequately, but appeared to be negligible.

4-14
 Figure 4-8. Typical Tube Inserts (Upper Left) and Plugs
(Lower Right)

A third tube leak repair method is used to protect the tubes against localized
inlet end erosion and/or corrosion. It consists of covering the tube 8 to
12 inches inward from the end with a coating such as coal tar epoxy or with a
metal or plastic tube insert. (See Figure 4-8 and 4-9.) It can be used only
to protect the tubes or to both protect and seal leaking tubes. It provides
effective protection and can substantially increase tube service life.

In a few special cases, condensers may develop leaks through the tube-to-tube-
sheet joint. Field surveys included two plants using rolled tube joints where
this type of leak was found. In one plant, the joint leakage was traced back
to poor workmanship during tube installation. In the other plant, the leakage
developed after circulating water hammer had severely overloaded the tube
joints and "pulled out" some of the tubes. In the first case, the leaks were
repaired for the most part by rerolling. In the second case, they were
repaired mostly by tube plugging.

4-15
Figure 4-9. Typical Tube Insert Setup

4-16
When numerous tube failures have occurred over some part or the whole of a
condenser's tube bundle, partial or complete tube replacement may be desirable.

A maintenance operation that is being used increasingly consists of testing the

tubes nondestructively from the water side. The eddy current method has
been used at several of the plants surveyed. This method can identify tubes
that have experienced major degradation and should be plugged as early as
practical. Also, the method may assist in defining the cause of the failures.
Eddy current findings may be of great value in pointing out potential prob­
lems, in allowing for efficient preventive tube plugging, and in supplying
data needed to forecast retubing requirements.

AIR INLEAKAGE

Modern power plants rely mainly on the condenser to collect noncondensible gases
and to provide a source point for their continuous removal. In fossil and PWR
plants, the noncondensible gases are mostly due to the leakage of air into com­
ponents that are below atmospheric pressure. The cycle components are built
and operated in such a way that they require very low oxygen levels on the
steam side (usually in the 5 to 15 ppb range). Some plants rely entirely on
the condenser for deaeration. Others include a feedwater deaerator that enables
operation with relatively higher condensate oxygen.

In BWR plants, most of the noncondensible gases are caused by the radiolytic
decomposition of water that takes place inside the nuclear reactor. Air
inleakage is experienced just as in any other plant, but its relative effect
on the cycle is normally quite small. BWR plants are also unique in the
following respects: the noncondensible gases are radioactive and can be
released to the atmosphere only after processing. Also, BWR cycle water
chemistry specifications require oxygen concentrations to be kept fairly high
(preferably between 50 and 100 ppb), and feedwater deaerators are never used.

In the following paragraphs the study findings of interest to air inleakage in

fossil and PWR plants are reported first. In many cases these findings are
also applicable to BWR plants. When required, additional information that
applies specifically to BWRs follows.

4-17
Air Leakage Impact

The field survey findings suggest that the impact of air inleakage on fossil
and PWR plants is sometimes underestimated. It was found that more than half
of the plants surveyed pay little attention to air inleakage as long as the
feedwater and condensate oxygen contents can be maintained within specifica­
tions during steady power operation. The findings also show that less than
half of the plants monitor the inleakage flow rate systematically and with
high accuracy.

A review of plant chemistry records shows that during startup and low load
operation, many units had relatively high oxygen and corrosion product levels
that could possibly contribute to steam generator fouling. This was informally
evaluated by two large utilities that attempted to correlate the pressure drop
buildup in fossil plant once-through boilers to the number of cold restarts.
The evaluation showed that for the units evaluated, boiler fouling was mostly
due to startups rather than long-term, relatively clean power operation.

The field survey also confirmed that copper alloy corrosion products are
potentially detrimental to all types of steam generators (fossil fueled, PWRs,
and BWRs). They may induce the following effects: 1) the alloy corrosion pro­
ducts themselves may adversly affect the boiler; 2) they may trap other water
impurities, such as iron corrosion products, and combine with them to impact on
the steam generator (heat transfer, pressure drop, and corrosion); and 3) they
may hamper the steam generator cleaning operation. It is well known that copper
alloy corrosion product release is strongly dependent upon cycle steam and
water oxygen concentration. Also, it is dependent on the concentration of
ammonia which is used for pH control in fossil-fueled and PWR plants. Air
inleakage has therefore a special effect on plants that use large amounts of
copper alloy heat transfer surfaces in their condenser and/or low pressure feed-
water heaters. This impact is further evaluated in Section 5, Performance of
Condenser Tube Alloys.

In BWR plants, the surveys show that air inleakage is relatively well monitored
because of its impact on the off-gas processing system. During normal opera­
tion, the direct effect of air inleakage on the cycle is essentially negligible.
This may be illustrated by considering that in a large BWR plant (rated at about
1100- to 1200-MWe), the condenser noncondensible radiolytic gas flow rate is in
the range of 250 scfm. In the same plant, only 5 to 10 scfm is due to air
inleakage.

4-18
Air Leakage Frequency, Size, and Duration

Air leakage occurs in a very different manner than does circulating water leak­
age. A plant can be made essentially watertight, but it cannot be fully pro­
tected against air inleakage. This is because of the many potential leak paths
around the active components maintained under vacuum (mostly valve stems and
pump shafts). As a result, plants always have some air inleakage which may be
classified as normal. In well-built and operated plants, the normal leak rate
is typically quite small. For example, the normal leak rate at a recently
completed 1100-MWe nuclear plant was about 3 scfm. The surveys also show that
many plants are continuously operated with air inleakage flow rates that
are relatively much greater. For example, most of the 600- to 900-MWe fossil
fueled units surveyed reported normal leak rates in the range of 15 to 30 scfm.

In addition, the following air leakage problems have been reported:

• During startup and low load operation, normal leak rates are
increased, sometimes by as much as a factor of 2 to 3 because
a greater part of the steam cycle is under vacuum.
• During periods between maintenance, the normal leak rates also
increase gradually, as valve stems and pump shaft packings
deteriorate because of wear or aging.
m Occasionally a component failure is experienced and a step
increase in leak rate results. The surveys indicate that
such failures are relatively rare. They occur no more than
once a year at any one plant on the average. These failures
sometimes result in a severe loss of condenser vacuum and thus
force an immediate plant shutdown.

Air Leakage Detection

The most straightforward air inleakage monitoring method is the direct measure­
ment of the flow rate of noncondensibles being removed from the condenser. In
fossil fueled and PWR plants, this measurement can be performed directly at the
outlet of the steam jet air ejector or vacuum pump. The measurement can be used
directly for trend monitoring. It can also be readily converted into an
absolute value by a temperature correction. In BWR plants, it is usually per­
formed at the off-gas system outlet point and is a reliable monitoring para­
meter during steady power operation. However, the measurement is difficult to
interpret during transients, such as startups, because the off-gas system
front-end equipment is then operating at very low yet fast-changing flow rates.

4-19
The direct measurement monitoring method is not always relied upon in fossil
and PWR plants. In those cases, the following two parameters may be monitored
for detection of potential gross air inleakage.

• Condenser backpressure is measured and correlated to plant

load and circulating water temperature. Abnormally high
readings are corrected to account for the estimated tube
fouling and interpreted in terms of potential condenser air
binding. Wherever needed, a pressure-to-flow-rate correla­
tion is made using the air removal system characteristics.
The method is obviously rather complicated and relatively
inaccurate, particularly because it relies in part on a
tube fouling estimate. Also, the correlation is based on
generally poorly shaped characteristics. It is used
because it is based on the backpressure monitoring that
any plant must perform to ensure good operating efficiency.
• The condensate oxygen concentration is also used as an air
inleakage indicator, again mainly because it must be monitored
to ensure satisfactory plant operation (in this case, corro­
sion protection). It is evaluated after correlation with
plant load, circulating water temperature, and estimated tube
fouling. Its interpretation, however, is probably even more
uncertain than that of backpressure, mostly because it
reflects only condenser air binding and thus is very sensi­
tive to local gas absorption and desorption phenomena that
take place below the air removal section. As it leaves this
section, the condensate is first subcooled and thus reabsorbs
noncondensibles (particularly during startup and low load
operation). Further down, however, the condensate is reheated
by rising steam that enters the tube bundle from below.
Figure 4-10 shows that moderate condensate temperature
changes (subcooling) can result in relatively large changes
in oxygen concentration. As a result, it is generally very
difficult to use the average condensate oxygen content as an
indicator of air inleakage. However, the method is effective
for detecting gross air inleakage through the condensate pump
shaft packing.

Air Leakage Location and Repair

Air leaks may be located during plant shutdown by filling the condenser steam
side with water—possibly dyed—and watching for potential outleakage around
suspected trouble areas such as valve stems and pipe joints. The field
surveys showed, however, that most air leaks are located with the condenser
under vacuum by the use of one or more of the following methods:

• Large leak detection by listening

0 Ultrasonic monitoring for medium to small leaks. Ultrasonic
detectors are very sensitive in low background noise areas,
and they were generally judged quite effective by all the
plants surveyed that had some experience with them.

4-20
 CONDENSER
TOTAL PRESSURE
inch of Hg
EQUILIBRIUM PARTIAL PRESSURE
OF NONCONDENSABLE, inch of Hg

CONDENSER
i; 4- TOTAL PRESSURE,
inch of Hg
EQUILIBRIUM CONDENSATE
0 9 CONTENT, ppm

2"
0.3 —

0.2 — 1"

5 10 15 20
SUBCOOLING °F

Figure 4-10. Equilibrium Noncondensible Partial Pressure and

Equilibrium Condensate Content from Air Inleakage
vs Condenser Total Pressure and Subcooling

4-21
 • Freon inleakage detection by monitoring the plant off-gas with
a halogen detector while dispersing freon in the vicinity of
the suspected leak points. The freon method is not as simple
as ultrasonic monitoring. However, some of the plants sur­
veyed found it to be the only one that could locate small
leaks in areas with high background noise.

Generally, the plants surveyed had only moderate success in permanently repair­
ing air leaks with the condenser under vacuum. Most plants reported using such
procedures as retightening packings or applying shellac around leaking valve
stems. These procedures usually result at least in a temporary reduction of
the leak. Permanent leak repair was reportedly performed mostly during plant
shutdown.

OTHER OPERATION AND MAINTENANCE CONSIDERATIONS

Satisfactory condenser performance requires consideration of several important

needs, limitations, and constraints not yet discussed. These additional con­
siderations are of a more general nature than the specifications reviewed
up to this point. For the most part, these considerations are function
oriented and are, therefore, presented under the functional headings listed
below:

• Startup
• On-line operation andmaintenance
• Off-line maintenance.

Startup

The startup of a condenser and its related systems raises two major concerns:

• Circulating water flow startup without excessive load tran­

sients on the condenser and pump
• Prompt turbine cycle cleanup to bring the condensate and feed-
water within operating water quality specifications

The following discussion considers the startup of plants using low-head circu­
lating water systems and the startup of cooling tower plants. For both kinds
of plants, the startup of the circulating water systems is reviewed first,
followed by a review of the turbine cycle cleanup.

4-22
Low-Head Circulating Water System Plants. Plants that discharge waste heat to
a lake, sea, or river generally use low head, mixed flow, high specific speed
circulating water pumps of the wet pit vertical type. The condenser is gener­
ally located well above the heat sink water level and a syphon effect is used
to reduce the pump head requirement. A typical circulating water system is
illustrated in Figure 4-11 which shows a twin shell, four water box condenser.
Pump characteristics are of the type illustrated in Figure 4-12. The circu­
lating water system is sometimes started with valves wide open, with no
special restriction or concern. Plants that require a long water supply line
between pumps and condensers must limit the initial circulating water flow
rate to protect against excessive water velocities and possible water hammer.
Great care must be applied to ensure that the pump flow rate is not throttled
down too far, as this could result in pump motor and thrust bearing loadings
greater than twice normal levels and also far greater than design levels. The
water boxes must be fully primed before the full flow can be established.

In some of the older, smaller units, these requirements are met in part by
connecting the condenser outlet water box to its discharging pipe at a point
higher than the tube bundle top tubes. The volume of air to be evacuated for
the syphon to be established is then quite small. In newer—and larger—plants,
a higher capacity water box air evacuation system is installed to allow for
direct priming well within the startup period. Essentially all units surveyed
as part of the present study fall in the second category and use a steam jet
air eductor for priming purposes.

During plant startup, the condenser contributes to the steam cycle cleanup in
the following ways: 1) it collects most of the cycle water and thus can be
used as a point source for the cycle treatment used to reduce dissolved and
suspended solid contents; and 2) it collects most of the cycle noncondensibles
and allows for their removal by the off-gas system. The first function is
straightforward and is performed readily, but the second one is difficult, par­
ticularly where the circulating water source is at very low temperature.
Several of the plants surveyed have been unable to maintain condensate oxygen
concentration within specifications during restarts. As indicated in the pre­
ceding Air Leakage Impact section, the study of the affected plant records
suggests that the problem may be more significant than is commonly recognized.
Excessive oxygen content appears to increase the amount of corrosion products
that are carried into the steam generators during startup to the point that
startups become the dominant source of external corrosion product input.

4-23
 A1

CONDENSER MAIN SYPHON

LEGS

WATER BOX
ISOLATION
VALVES

RETURN

WET PIT
VERTICAL
PUMPS STOP-CHECK
VALVES

INTAKE
STRUCTURE

Figure 4-11. Typical Circulating Water System for Twin Shell

Condenser with Open Cycle or Cooling Lake

Figure 4-12. Typical Characteristics of Low Head Circulating Water Pumps

4-24
 High-Head Circulating Water System Plants. Cooling tower plants generally use
mixed flow, medium head, moderate specific speed circulating water pumps of the
wet pit vertical type. The circulating water system is similar to that shown
in Figure 4-11 except that it includes a vertical upward leg to the tower on
the downstream side instead of a downward syphon leg. Also, cooling tower
systems generally have a tower bypass line (at least part flow). They also
include condenser downstream water box isolation valves to protect against down­
stream end positive pressure. The pump characteristics are of the type
illustrated in Figure 4-13. Water boxes are maintained at a positive pressure
during operation, and they are vented directly to the atmosphere.

%
ICY,
BRAKE HORSEPOWER

ALLOWABLE OPERATING RANGE

CIRCULATING WATER FLOW RATE % OF DESIGN

Figure 4-13. Typical Characteristics of Medium Head Circulating Water Pumps

The systems are normally started with the pump discharge valve throttled to
about 30 to 50 percent capacity, with the condenser isolation valves wide open
and with the tower bypass open. Bypass recirculation is maintained until the
circulating water temperature is increased to the point where effective conden­
sate deaeration is achieved.

Theoretically, the cooling tower plant circulating water system should be fairly
easy to start. The survey confirms however, that a number of such systems have
had the following startup problems:

• Water hammer, usually created by excessively fast or premature

opening of the pump discharge valve
• Water box overpressurization caused by excessively fast clos­
ing of the condenser outlet valves

Both of these examples resulted in severe condenser damage.

On-Line Operation and Maintenance

The surveys show that condenser on-line operation and maintenance requirements
vary over a very broad range, mostly as a function of circulating water compo­
sition. Most plants, however, perform on-line the functions listed below:

• Monitoring for

Water inleakage
Air inleakage
Heat transfer surface fouling

• Maintenance

Tube cleaning (either on-line or during part-load

operation)
Water leak location and repair (during part-load
operation)
Air leak location and repair

These functions have been reviewed in the preceding sections. The following
paragraphs provide some additional information on tube cleaning.

4-26
On-line tube cleaning operations can be separated according to their objectives
into the following two categories:

Removal of the small deposits—or fouling—that gradually

build up on the tube inner walls. Fouling may result in
reduction of heat transfer coefficients and/or accelerated
corrosion. The deposits may be caused by thermohydraulic
and chemical phenomena or biological growth.
Removal of the relatively large-sized foreign matter that is
mechanically entrained in the condenser. This matter may
result in complete or partial tube plugging and thus induce
tube erosion, corrosion, and fouling.

Biological Growth Control by Chemical Addition. This is the primary tube

cleaning method used at most of the plants surveyed (63 percent). It consists
of adding sodium hypochlorite or chlorine gas to the inlet circulating water in
such amounts that the outlet circulating water contains between 0.5 and 1 ppm
of equivalent residual free chlorine (now limited to 0.2 ppm average—0.5 ppm
maximum by 40 CFR 423). A few of the plants surveyed use or have used
biodegradable biocides instead of chlorine.

Frequency, duration, and input dosage used vary greatly from plant to plant and
from one season of the year to the other. For the most part, the chemical
addition requirements are a function of the circulating water composition and
temperature, of the condenser tube inherent toxicity, and of the effectiveness
of the other tube-cleaning methods used.

Biological Growth Control by Thermal Treatment. This method is used at a

number of seawater and brackish water-cooled plants. It consists of recircu­
lating the outlet circulating water back to the intake as required to raise
the circulating water system temperature to a level that is lethal for
dominant marine organisms (usually about 120°F). It has been found to be
effective even against large organisms such as clams and mussels which have a
high tolerance for chemicals. This method has also been found to be generally
more acceptable from an environmental viewpoint. The major drawback is the
relatively high initial cost of the recirculating line or channel that is
required.

Circulating Water Chemical Treatment. This method is used at a number of

closed cycle freshwater-cooled plants that are operated on cooling towers or
dedicated cooling lakes. It generally includes the addition of sulfuric acid

4-27
to prevent the precipitation of calcium carbonate. At plants that operate with
a high calcium sulfate concentration product, a dispersing agent is employed.
Two of the plants included in the field survey use such an agent and have
reported it to be very beneficial in reducing deposit buildup.

On-line Mechanical Cleaning. This method is available at 12 percent of the

plants surveyed. Two such cleaning systems have been developed, MAN and
Amertap. No MAN system was installed in the plants surveyed. The surveys
show that the Amertap system is very effective in removing all deposits,
with the possible exception of the relatively hard calcium sulfate scale for
which system effectiveness could not be adequately confirmed. The Amertap
plants surveyed used relatively old system designs and were not equipped
with the now frequently recommended debris filter. Many plants reported
excessive ball losses apparently due to the buildup of debris in the outlet
ball catcher and the spurious ball catcher flap opening caused by the resulting
high pressure drop. Because of this, a number of plants have discontinued
system operation. The problem could not be fully evaluated; however, it
appeared that most ball losses would be eliminated by the upstream debris
filter. A second problem reported with some of the systems is ball hide-out.
For example, one of the plants surveyed reported short term hide-outs of up to
several thousand balls. This second phenomenon is unusual, and a cure has
apparently not yet been firmly established. It probably results, at least in
part, from a combination of characteristics of the circulating water system
(dead pockets - low velocity regions) with a mismatch of the ball weight to
actual buoyancy. It appears that correction of the mismatch could reduce the
problem to acceptable levels. Further data are needed, however, to confirm this
hypothesis.

Reversing the Direction of Flow. Another on-line mechanical cleaning method

consists of reversing the direction of flow across the condenser. It is used
by approximately 3 percent of the plants included in the survey. It is aimed
at backwashing the miscellaneous debris, grass, and leaves that accumulate on
the inlet tubesheet. It is very effective at doing this, but its potential is
limited and its initial cost is fairly high.

4-28
Off-line Maintenance

Condenser off-line maintenance requirements also vary over a broad range,

mostly as a function of condenser status, tube material, and circulating water
composition. Most plants systematically perform some of the following
maintenance tasks:

• Inspect the condenser steam side once a year, paying partic­

ular attention to 1) indications of tube vibration and/or
steam impingement, 2) the condition of insulation lagging,
3) the condition of drain or dump line inlet baffles, and
4) the presence of loose parts. At the same time, check the
condenser and attachments for possible small air and water
leaks; update the tube plugging records; confirm the suit­
ability of the guard plugging schemes used, if any; inspect
a sample of tubes on the water side for possible fouling,
corrosion, and/or erosion; and check the condition of in­
stalled tube plugs and generally inspect the condenser and
its related systems
• Supplement the above by performing eddy current examination
of tubes, as described in the Water Leakage Repair section
(see page 4-17)
• Protect the tube inlet end with coatings and/or inserts
® Maintain, as required, the tubesheet and water box wall coat­
ing and the cathodic protection electrodes and equipment
(see Figure 4-14)
• Periodically perform manual tube cleaning to supplement, as
required, the on-line cleaning method(s) and restore efficient
heat transfer

Sixteen percent of the plants surveyed reported manual mechanical cleaning as

their primary cleaning method. In most cases, cleaning is performed by insert­
ing and shooting through each tube some kind of tight-fitted element such as an
Amertap ball, a MAN-type of cylindrical brush, or a plastic, rubber, or sheet
metal scraper (see Figure 4-15) . Some plants reported using high pressure
traveling water jets that are run along the whole length of the tubes, other
plants perform a high velocity water flush, and a few plants used tube rodding.

Overall, manual mechanical-type cleaning methods were reported to be quite

effective in restoring tube cleanliness. However, they are very time consuming
and can be relied upon only when the tube fouling rate is fairly low. Generally,
these methods are implemented during cold shutdowns, but some plants reported
using them during off-peak hours, after partial load reduction and drainage of
one condenser section.

4-29
Figure 4-14. Typical Cathodic Protection Electrodes

Figure 4-15. Typical Tube Cleaning Devices

 Section 5

PERFORMANCE OF CONDENSER TUBE ALLOYS

INTRODUCTION

This Section contains the study results that are related to the performance
of different condenser tube materials in various service environments. The
classifications used to define the exposure conditions and the concepts used
for presentation of the results are discussed. The survey findings are
presented by tube alloy groups. A comparative summary of the performance of
the commonly used alloys concludes the section.

Service Environment Classifications

Condenser tubes fail as the result of many different mechanisms. To evaluate

their performance requires an orderly classification of the several service
parameters so that comparisons can be made with equivalent data. A logical
consideration of the various failure mechanisms leads to the following broad,
categorical parameters which define the service environment.

• Tube material, dimensions, and physical properties

• Service sector within the condenser
• Steam-side environment
• Water-side environment
• Failure mode

These categories were used for the organization and evaluation of the survey
data. Each is discussed in more detail below.

Tube Material, Dimensions, and Physical Properties. The material composition

of the condenser tube is one of the major variables that affects its service
performance. Accordingly, survey results are segregated by individual alloy.
It was, however, necessary to lump all admiralty brasses into a single category
because the admiralty alloy used was not known at several stations.

5-1
Tube diameter was found to have no discernible effect on performance, probably
because of the narrow range of diameters in use. For a given alloy, the tube
wall thickness is usually the same; e.g., all admiralty tubes surveyed were
18 BWG. For other alloys, wall thickness variations were not more than one
gauge. While condenser tube lengths vary considerably, a preliminary
analysis of the survey data indicated no correlation between tube length and
failure incidence. Hence, tube dimensions are not of prime concern in the
assessment of condenser tube failures.

The physical properties of the tubing material are important with regard to
susceptibility to erosion and, to a degree, vibration. But because physical
properties are specific for a given alloy and gauge, they need not be classi­
fied separately.

Service Sector in the Condenser. It is apparent that tubes in the three main
service sectors of the condenser are subject to different failure mechanisms
and failure rates. Tubes in the outer few rows (impingement section) are more
susceptible to failure due to erosion by impinging steam and to severance or
puncture by missiles or broken baffles than are the tubes in the main condens­
ing section. Likewise, if the tubes in or immediately adjacent to the air
removal section are of an ammonia-sensitive alloy, they can be expected to
suffer a higher incidence of ammonia induced failures. This is because
localized concentrations of ammonia can be orders of magnitude greater than
the ammonia concentration in the bulk condensate stream.

In most condensers with admiralty tubes in the condensing section, tubes of a

harder material such as stainless steel or copper-nickel are used in the
impingement area. The incidence of ammonia attack may also be reduced by the
use of stainless steel or copper-nickel in the air removal section.

Obviously, in a condenser containing tubes of more than one material, all

tubes cannot be lumped together for failure-rate evaluation. It is also
desirable to segregate the three service areas of the condenser if valid
comparisons are to be made. Therefore, all condensers surveyed were sub­
divided into several "tubesets." A tubeset was defined as a group of tubes

5-2
of the same material, service sector, and age, for which failure rate infor­
mation was available. The following tubeset designations were used:

A = air removal section

I = impingement section

C = main condensing section

0 = all tubes in shell (must be of same material) equivalent to

A + I + C for shell and used only when individual data were
not available

B = complete bundle within shell (same qualifications as in 0)

X = C+ I when of same material and individual data not available

Y = I + A when of same material and individual data not available

Z = C + A when of same material and individual data not available

S = complete shell, combination of subsectionsfor which individual

data were available, may contain more than one material (these
are repeat data and were excluded from most calculations)

In order to simplify the manipulation of the data and to give the largest
possible data base, the categories that represent primarily condensing section
tubes were combined:

M (main body) =C+0+B+X+Z

Both the individual and combined designations are tabulated in the Field Survey
Matrix.

The tubeset concept was applied both to the existing and the replaced sets of
tube s.

For example, a single condenser with two shells, each containing two bundles
and having a separate air removal section in each bundle, would be divided into
twelve tubesets. The tubesets would consist of the air removal, impingement,
and condensing sections for each of the four bundles. On the other hand, if
tabular data were available which listed only the total number of plugs in each
bundle and no plug location drawing was available, then the condenser could be
subdivided only into four-bundle tubesets — provided all the tubes were of the
same material. If the condenser had been retubedf then whenever possible, data
were obtained for the replaced tubesets.

5-3
Steam-Side Environment. The steam-side environment can be responsible for tube
failures from several mechanisms. Mechanical failures due to missiles, broken
baffles, etc, occur on the steam side only. Other failures are caused by steam
or water droplet impingement/erosion. To some degree, failures due to tube
vibration relate to steam-side operating conditions. Most steam-side
failures from electrochemical corrosion are induced by localized high concen­
trations of ammonia. Accordingly, the field survey sought all available
information relevant to the steam-side environment. This included data on
turbine cycle chemistry procedures such as chemicals fed, dosage, control
methods and criteria, as well as analytical results both present and over the
lifetime of the tubeset.

Water-Side Environment. The water-side environment is the most important

parameter relative to tube failure frequency. While the chemistry and biology
of the circulating water apparently are responsible for most water-side
failures, other factors are also significant. Cavitation/erosion due to air
release causes many tube failures as does erosion caused by high velocities in
regions of partial tube blockage. Accordingly, data were sought not only on
the chemical, physical, and biological properties of the circulating water, but
also on water box configuration and the hydraulics of the circulating water
system.

The cooling water at each station was classified into one of the following nine
categories:

R = River
L = Lake
M = Polluted freshwater
T = Cooling tower
S = Recycled spray pond or canal with concentrating mechanism
B = Brackish — variable from fresh to salty
K - Polluted brackish
C = Clean seawater
P = Polluted seawater

These classifications are given in column 14 of the Field Survey Matrix.

Because the various polluted categories were based primarily on the subjective
opinion of station personnel rather than analytical data, and to facilitate

5-4
data manipulation with the largest possible arrays, the above categories were
subsumed as follows:

Water Type Designation Original Categories Included

Fresh F R, L, M
Ocean 0 B, K, C, P
Concentrated J T, S

The combined water categories are tabulated in column 171 of the Field Survey
Matrix.

The distribution of tubesets by tube alloy and cooling water category is shown
in Table 5-1.

Table 5-1

DISTRIBUTION OF TUBESETS BY
MATERIAL AND WATER TYPE

ALLOY TUBESETS IN TUBESETS IN TUBESETS IN TOTAL

FRESHWATER TOWER WATER OCEAN WATER
NO. % NO. % NO. % NO. %

Admiralty Brass 134 45 18 31 1 <1 153 26

90-10 Cu-Ni 20 7 6 11 58 26 84 15
70-30 Cu-Ni 6 2 12 21 57 25 75 13
Cu-Ni (specific type 4 1 0 0 0 0 4 <1
unknown)
Aluminum Brass 0 0 0 0 52 23 52 9
Aluminum Bronze 0 0 0 0 12 5 12 2
Type 304 Stainless 117 39 21 37 0 0 138 24
Steel
Type 316 Stainless 8 3 0 0 23 10 31 5
Steel
Titanium 8 3 0 0 23 10 31 5
Total 297 100 57 100 226 100 580 100

Failure Mode. From the onset of the study it was evident that identification
of the causative failure mechanism would be possible only in rare instances.
Often such determination cannot be made until after the failed tube is removed
and submitted to comprehensive analysis and, on occasion, not even then. None­
theless it was desirable to classify failure causes in the most practical manner

5-5
based on the data that could be gathered during the relatively brief field
surveys. Helpful information would be gained if the reason for plugging the
tubes could be placed in the following categories:

1. Mechanical failures including missiles, vibration, severance, etc.

2. Failures originating on the steam side for any reason other than
mechanical.
3. Failures originating on the water side for any reason whatsoever.
4. Failures around the tube-to-tubesheet joints.
5. Tubes plugged to guard against anticipated early failure. At
many plants tubes adjacent to failed tubes will be plugged
because their imminent failure is anticipated.
6. Total failures for all reasons where any tube plugged was con­
sidered to have "failed."

Accurate information on Total Failures was obtained at all surveyed stations.

In most cases the location of the plugs was accurately known as was the date
when the plugs were installed. Information that allowed distribution of the
plugs into the other five failure mode categories was available far less often
than expected.

Failures classified into categories 1, 4, and 5 are virtually independent of

the tube material. Failures classified into categories 2 and 3 are usually a
function of the specific tube alloy. The 6th category — total tubes plugged
— is obviously a mixed grouping which does not allow ready interpretation of
failure mode or the location of the failure within the tube or the tube within
the tubeset. However, it is the only category for which data were always avail­
able at the surveyed stations and, of necessity, most of the results presented
in this section are based on the Total Plugged category. Actually, performance
results calculated from the Total Plugged category are a practical measure of
condenser tube serviceability because condenser performance and maintenance
costs are affected by the number of tubes plugged, not by the reason for the
plugging.

Performance Rating Factor

Quantitative analysis of the performance of condenser tubes of different alloys

in various environments requires the development of a suitable mathematical
expression for the rate of failure. Yet, because of the many different failure
modes, a single expression that precisely describes the total failures versus
time curve is not possible. The number of tubes failed as a function of time

5-6
can be expected to increase exponentially for certain classic corrosion modes
of failure. But failures due to missiles, broken baffles, etc, are independent
of time. Over the years, the power industry has developed a number of methods
for analyzing and interpreting failure rates.

After a review of the problem it was decided that the survey results would be
quantified by the use of the Newton and Birkett (46) formula:

F/N x 100
_4
Service Hours x 10

where FR = Failure rate, percent of tubes plugged per 10,000 hours

of operation
F = Number of tubes failed
N = Number of tubes in tubeset.

This expression yields the linear or average failure rate during the time
interval from the first service of the tubeset to the time of calculation. The
value calculated for any single time, in most instances, does not describe the
true shape of the F vs t curve over the selected time interval. The failure
AF dF
rate /At is numerically equal to /dt only when the F vs t curve is linear.
If sufficient tubeset data are available to enable calculation of FR at frequent
intervals through the service life of the tubeset, an accurate and adequate
description of the failure curve results. Unfortunately, ample data were not
always available for the surveyed tubesets.

The calculation sheet used to tabulate the several failure rates is shown in
Table 5-2. To the extent the collected data allowed, the failure rates were
calculated for each failure mode (FRM, FRS, FRW, FRJ, FRG, and FR) at accumu­
lated service hour periods of 10,000, 20,000, 40,000, 60,000, 90,000, 120,000,
and 180,000, as appropriate, up to the survey date or for the end of service of
the replaced tubesets. For relatively new tubesets (less than 20,000 hours in
service), the time periods used were 1/10 of those indicated. The distribution
of tubesets for which the total failure rate (FR) could be calculated is shown
in Table 5-3 by tube material, condenser service sector, and cooling water type.
The distribution of the total number of condenser tubes and the condensing
service area is shown by tube alloy and cooling water category in Table 5-4.

5-7
 Table 5-2

EPRI CONDENSER LEAKAGE STUDY - JOB 11540-001

TUBE SET FAILURE RATE CALCULATION SHEET

STATION_______ UNIT_____ _ SHELL_______ BUNDLE_______ SET_

TUBE SET NO.__________ TUBE SET DESCRIPTION_______________
TUBE MATERIAL_______ O.D. _____ GAGE BWG LENGTH^________
TUBE MANUFACTURER _______ __ NO. OF TUBES IN TUBESET
INITIAL SERVICE DATE / / TOTAL TUBE SET LIFE ________
PRESENT TUBE SET?

ACTUAL
OPERATING HOURS
DATE / / / / / / / / / / / / / / / /

CAUSE STANDARD OPERATING INTERVALS (X10 3 HRS)

OF FAILURES TO DATE OR
10 20 40 60 90 120 180
END OF LIFE
MECHANICAL
(FRM)
PLUGS
FR
STEAMSIDE
(FRS)
PLUGS
FR
WATERSIDE
(FRW)
PLUGS
FR
JOINT
(FRJ)
PLUGS
FR
GUARD
(FRG)
PLUGS
FR
TOTAL
(FR)
PLUGS
FR
NOTES: PLUGS
NO. OF TUBES x 106
OPERATING HOURS

5-8
 1

Table 5-3
DISTRIBUTION OF TUBESETS WITH KNOWN FAILURE RATES
BY MATERIAL, CONDENSER SERVICE, AND WATER TYPE

MATERIAL TUBESETS IN TUBESETS IN TUBESETS IN

FRESH WATER TOWER WATER OCEAN WATER TOTAL
M A I TOTAL M A I TOTAL M A I Y TOTAL M A I Y ALL %
NO. % NO. % NO. %

Admiralty 66 34 2 102 44 16 0 0 16 36 1 0 0 0 1 <1 83 34 2 0 119 25

90-10 CuNi 16 1 0 17 7 0 4 0 4 9 55 1 0 0 56 29 71 6 0 0 77 16

70-30 CuNi 0 0 0 0 0 2 10 0 12 27 6 45 0 2 53 28 8 55 0 2 65 14

Specific
CuNi, Type 0 0 4 4 2 0 0 0 0 0 0 0 0 0 0 0 0 0 4 0 4 21
Unknown
Aluminum
Brass 0 0 0 0 0 0 0 0 0 0 35 2 0 0 37 19 35 2 0 0 37 8
Aluminum
Bronze 0 0 0 0 0 0 0 0 0 0 12 0 0 0 12 6 12 0 0 0 12 3

Type 304
Stainless 26 42 27 95 41 9 2 2 13 29 0 0 0 0 0 0 35 44 29 0 108 23
Steel
Type 316
Stainless 0 4 4 8 3 0 0 0 0 0 2 3 4 0 9 5 2 7 8 0 17 4
Steel
Titanium
Total 0 4 4 8 3 0 0 0 0 0 21 2 0 0 23 12 21 6 4 0 31 7
TOTAL 108 85 41 234 27 16 2 45 132 53 4 2 191 267 154 47 2 470

M - Main Body
A - Air Removal Section
I - Impingement
Y - I + A
 Table 5-4

DISTRIBUTION OF NUMBER OF TUBES

AND CONDENSING SURFACE AREA
BY MATERIAL AND WATER TYPE

ALLOY FRESH WATER TOWER WATER OCEAN WATER PERCENT OF TOTAL

NO. SURFACE NO. SURFACE NO. SURFACE NUMBER SURFACE
TUBES AREA TUBES AREA TUBES AREA TUBES AREA
xl03 pji2 xlO3 FT2 xlO3 FT2
Admiralty 499,534 4,779 144,632 1,399 5,340 43 29 26
90-10 Cu-Ni 74,540 864 2,008 20 422,635 4,601 22 24
70-30 Cu-Ni 1,208 10 23,404 238 79,484 882 3 5
Cu-Ni (Spe­
cific type 512 4 0 0 0 0 <1 <1
unknown)
Aluminum
Brass 0 0 0 0 333,203 3,345 15 14
Aluminum
Bronze 0 0 0 0 88,550 814 4 4
Type 304
Stainless 230,160 2,151 174,968 1,781 0 0 18 17
Steel
Type 316
Stainless 1,720 16 0 0 37,605 469 2
Steel
Titanium 1,284 14 0 0 163,400 1,690 7 7
Misc 7,212 69 0 0 0 0 <1 <1
Total 816,170 7,907 345,012 3,438 1,130,217 11,844 100 100

Probability Calculation

A practical and useful comparison of failure rates for different materials and
service conditions can be obtained by calculating the probability that a given
failure rate will be attained. This was done whenever the field survey yielded
sufficient data. The probability was calculated after Newton and Birkett (46)
as shown in the following example:

The data bank for admiralty tubes in main body service in fresh water consisted
of 66 tubesets for which the total terminal failure rate could be calculated.
The calculated failure rates (FR) were arrayed in numerical order and ranged
from 0.000 to 2.565. The median FR was 0.031, the arithmetic average of the

5-10
thirty-third and thirty-fourth value in the array. Fifty percent of the tube-
sets had a failure rate less than 0.031; hence there is a 50-percent probability
that the FR will not exceed 0.031. In 6 of the 66 cases, the failure rate
exceeded 0.333; hence the probability of an FR of 0.333 or less was 91 percent.
Likewise, 58 of the 66 tubesets had an FR of 0.222 or less; hence, the proba­
bility is 88 percent that an FR less than 0.222 will be attained.

The validity of the calculated probability values is clearly dependent on the

number of tubesets included in the data bank and how representative they are.

The calculated probabilities can be used for approximate estimates of the

anticipated service life of a tubeset. If a tubeset served 7,500 hours a
year for 40 years and at that time had 10 percent of the tubes plugged, the
failure rate would be 0.333. Based on the data collected in this survey (given
in the example above) for admiralty tubes in main-body service with fresh
cooling water, there is a 91 percent probability that the failure rate will be
0.333 or less. This corresponds to 40 years of service with not more than
10 percent of the tubes plugged. Also, the probability is 88 percent that
admiralty tubes in main body service in fresh water will have a failure rate
less than 0.222, which corresponds to 30 years of service with not more than
5 percent of the tubes plugged.

Obviously, caution must be exercized in the application of this estimated

service life especially for data arrays that contain relatively few, or
relatively young, tubesets. Inherent in this calculation procedure is the
assumption that the attained failure rate is independent of the age of the
tubeset. This assumption is verified for the data collected in this survey
which show no correlation between the total failure rate and the age of the
tubeset. Whether this assumption will be valid over 40 years for titanium,
for example, where the oldest tubeset has been in service only 6 years cannot
be affirmed from the data now available.

Presentation of Results

Most of the study results are presented in plots of Failure Rate versus
Probability calculated as described above. Most of the graphs are in two
sections. The upper section is a graphic nomograph that enables rapid
determination of the failure rate that corresponds to a given service life in
hours (or years at 7,500 hours per year), and a given percentage of plugged

5-11
tubes. The curve plotted in the lower section shows the probability (as
calculated from the particular data base) that the corresponding failure rate
will not be exceeded. Within the limitations described, these plots are
believed to be valid and useful presentations of the historical data on con­
denser tube performance as collected by this survey. While individual
exceptions to the trends shown must be expected, in view of the breadth and
depth of the survey sample as described in Section 3, the trends should be
typical of industry-wide experience.

COPPER-BASE ALLOYS

Copper-base alloys have historically been used for condenser tubing materials
because of their good heat transfer coefficients, anti-biofouling charac­
teristics, and relatively good corrosion resistance. Table 5-1 shows that of
the total 580 tubesets surveyed, 380 or 66 percent were copper-base alloys.
Table 5-4 shows that 73 percent of the tubes and 73 percent of the total con­
densing surface area in the surveyed plants were fabricated from copper-base
alloys. Table 5-5 lists the copper-base alloys found in the surveyed con­
densers, the copper alloy number (CA), the designation used in this study,
and the chemical composition of each alloy. Note that no distinction is made
between cooper alloys 443, 444, and 445, the admiralty brasses. The plants
surveyed often did not know which of these alloys was in service. Furthermore,
the performance of each is similiar.

Historically, admiralty brass condenser tubing has been used for freshwater
service while the copper-nickels, aluminum brass, and aluminum bronze are used
predominantly in seawater cooled plants. This trend can be seen in Table 5-4.
In the past decade, the copper-nickels, and especially 90-10 Cu-Ni, have been
specified more often as tube material for seawater cooled plants. The copper-
nickels, i.e., 70-30 Cu-Ni and 90-10 Cu-Ni, have lower corrosion rates in sea­
water and are more resistant to ammonia, attack than the other copper-base
alloys. All of the copper alloys are highly susceptible to erosion-corrosion.
Erosion-corrosion includes degradation caused by air release and locally
turbulent, high velocity water.

A chromium bearing copper-nickel alloy, IN838 (CA722), has been developed which
reportedly has significantly increased resistance to erosion-corrosion in sea­
water. However, no condensers have yet been tubed with CA722 and thus no oper­
ating data are available.

5-12
 1

Table 5-5
COPPER-BASE ALLOYS
CHEMICAL COMPOSITION AND DESIGNATION

STUDY COPPER
DESIG­ ALLOY ALU­ LEAD, MANGA­ ANTI­ PHOS­
MATERIAL (COMMON NAME) NATION NUMBER COPPER TIN MINUM NICKEL MAX. IRON ZINC NESE ARSENIC MONY PHORUS

0.9- 0.6 0.02-

Admiralty Brass AD 443 70-73 1.2 _ . 0.07 max R* 0.10 _

0.9- 0.6 0.02-

-13

Admiralty Brass AD 444 70-73 1.2 _ - 0.07 max R* - - 0.10 -

0.9- 0.6 0.02-

Admiralty Brass AD 445 70-73 1.2 _ 0.07 max R* - _ 0.10
86.5 9.0 1.0- 1.0 1.0
90-10 Copper-Nickel CL 706 min - - 11.0 0.05 1.8 max max - - -

65.0 29.0- 0.40- 1.0 1.0

70-30 Copper-Nickel CN 715 min _ _ 33.0 0.05 0.70 max max - - -

76.0 1.8- 0.06 0.02-

Aluminum Brass BA 687 79.0 _ 2.5 - 0.07 max R* _ 0.10 - _

93.0 5.0- 0.10 0.02-

_ _ _
Aluminum Bronze AB 608 min _ 6.5 _
0.10 max 0.35 _

*R = Remainder
Admiralty Brass, AD; CA 443, 444, 445

Freshwater Service. Admiralty brass was the most common tube alloy for fresh­
water service in the surveyed condensers. It accounted for 44 percent or 134
of the tubesets, 62 percent of the tubes, and 61 percent of total condensing
area. Figure 5-1 indicates that the 66 admiralty brass tubesets in main-body
service have an excellent performance record in fresh cooling waters. There
appears to be a greater than 90-percent probability that admiralty tubes in
main-body service will last 40 years with less than 10 percent of the tubes
plugged. There also appears to be at least 90-percent probability that
admiralty brass will provide 40 years service with only 5 percent of the tubes
plugged. The median failure rate was 0.031. Only two of the 66 admiralty tube-
sets in this service had been replaced.

Contributing failure mechanisms were given for many of the tubesets. Some
tubesets had more than one failure mechanism given; others had none. In order
to gain some insight on the impact of a particular mechanism, the number of
times each failure mechanism was cited was tabulated. Table 5-6 indicates that
waterside corrosion accounts for most of the failure mechanisms attributed to
admiralty brass in freshwater. Most of these failures are in the "general
corrosion — specific type unknown" or "stress corrosion cracking" categories.

Table 5-6

CONTRIBUTING FAILURE MECHANISMS

FOR ADMIRALTY BRASS IN FRESH WATER

AIR REMOVAL TUBESETS ALL TUBESETS

No. Of Times No. of Times

Cause Cited Percent Cited Percent

General corrosion or
specific type unknown 26 74 36 38

Stress corrosion cracking 8 23 14 15

Pitting
Erosion-corrosion 1 3 11 11
5 5
Falling objects 2 2
Tube joints 4 4
Steam erosion 10 10
Vibration 7 7
Mechanical 7 7
TOTAL 35 100 96 99

5-14
 wm

PERCENT
PLUGGED

40 YEARS

Ul
i
H
Ul

0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
FAILURE RATE

Figure 5-1. Failure Rate vs Probability Curves for Admiralty Brass in Freshwater
Service. Curves for Main-Body Service (M) (66 Tubesets) and the Air
Removal Section (A) (34 Tubesets) are Shown.
Cooling Tower Service. Failure rate data were available for only sixteen tube-
sets in cooling tower service and one tubeset in ocean water service. Since
the data base was so small, only limited performance evaluations were attempted.
The observed median failure rate was 0.135, over 4 times greater than for tube-
sets in freshwater service.

Effect of Ammonia. Admiralty brass is highly susceptible to ammonia attack, and

also to ammonia-induced stress corrosion cracking. High concentrations of
ammonia can exist in the air removal sections of surface condensers. Wilson
(32), for example, has estimated the ammonia level in the air removal section
of a condenser to be as high as 125 ppm even when there is only 0.2 ppm in the
hotwell. Andres noted concentrations of approximately 400 ppm in condensate
sampled adjacent to the tubesheet in the air removal zone with ammonia feed-
water treatment controlled at 0.2 ppm.* He stated that adding morpholine to
the feedwater will increase the ammonia concentration to 1000 ppm (due to
decomposition) at the same location. Coit, et al,.gave an example where a
utility collected condensate from tubes in the air removal section and found
it to contain as high as 500 ppm ammonia (189). Fink and Boyd reported the
same value.** Aoyama el al-(185) observed 45 to 2700 ppm ammonia in the steam
and 4 to 258 ppm ammonia in the condensate. In view of the above, admiralty
brass in the air removal zone would be expected to perform poorly. This was
found to be the case in the surveyed condensers and is best shown in Figure 5-1.
Using the 10-percent of tubes plugged criterion, it can be seen that there is
less than 30-percent probability that admiralty brass in the air removal sec­
tion will attain a 40-year life.

Effect of Hotwell Ammonia Concentration and Shrouding. Shrouded air removal

sections may contain higher concentrations of ammonia than unshrouded ones.
Therefore, an attempt was made to correlate failure rate and hotwell ammonia
concentration for shrouded and unshrouded sections. The correlation attempt
was performed for all admiralty tubesets and for those used only in air removal
service. There appeared to be no correlation between hotwell ammonia concen­
tration and failure rate.

*Andres, R.F., Proceedings of the 18th Water Conference, 1957, Discussion p. 102.
**Fink, F.W. and Boyd, W.K., Topical Report on "Corrosion Problems in Power Plant
Steam Condensers," to the Copper Development Association, May 1971, Battelle
Columbus Laboratories ,

5-16
In order to determine the effect of shrouding, the ratio of shrouded to
unshrouded tubesets (Rs) was calculated for several failure rate increments
For all the tubesets studies this ratio was 1.32. Except for the very low
failures rates and very high failure rates, i.e., those less than 0.01 and
greater than 10.0, respectively, Rs increased as the failure rate increased
as shown in Table 5-7.

Table 5-7

RELATIONSHIP BETWEEN THE SHROUDED-

UNSHROUDED RATIO, Rs, AND FAILURE RATE

FAILURE RATE INCREMENTS

<0.01 0.01 to 0.1 0.1 to 1.0 1.0 to 10.0 >10.0

RS 3.33 0.50 2.29 3.25 1.17

Number of
Tubesets 13 34 21 17 13

It was also noted that Rs equaled 1.00 for failure rates less than 0.33 (that
failure rate which is equivalent to 10-percent plugged condenser tube, after
300,000 hours or 40 years service) and 2.08 for failure rates greater than
0.33. There appears then to be some correlation between high failure rates
and shrouded air removal sections. This correlation may have been greater
if sufficient data were available for determining failure rates caused only by
ammonia concentrations, rather than the total failure rates which were used.

Effect of Condenser Size. An attempt was also made to correlate the failure
rate of admiralty tubes and condenser size. Failure rate versus total conden­
sing area was plotted and the condenser manufacturer was noted for each data
point. However, no correlation was evident and no attempt was made to do this
correlation for other condenser tube materials.

Effect of Operating Hours. An attempt was also made to correlate the total
failure rate (all causes) and the total operating hours for admiralty brass
in freshwater service. The raw data were plotted (i.e., failure rate vs the
time increments given in the Rating Factor subsection) for each of fourteen
stations using admiralty brass in freshwater. Separate curves were plotted
for air removal tubesets which have been retubed, air removal tubesets now

5-17
in use, main-body tubesets that have been retubed, and main-body tubesets now
in use. In most cases, the failure rate decreased with time (i.e., the time
increments) after an initial increase from zero. In one-third of the remaining
cases, the rate of failure rate increase (i.e., the slope of the failure rate
vs time curve) decreased with time. The curves for the air removal tubesets
(which, in the case of admiralty brass, can be associated with steam-side
corrosion failures) and the main-body tubesets were qualitatively identical.

In general, there appeared to be no correlation between failure rate and the

time increments and no attempt at this correlation was made for other tube
materials. This does not necessarily mean that there is no correlation
between failure rate and time. It is obvious that failure rate will decrease
with time when only mechanical failures (e.g., falling objects) are considered.
Typically, a large number of tubes will be plugged immediately following a
mechanical failure and very few will be plugged later (as a result of this
failure). On the other hand, an increase of failure rate with time for some
corrosion-related failures can be postulated and was indeed observed at some
plants.

90-10 Copper-Nickel, CL, CA706

Freshwater Service. Although 90-10 Cu-Ni is primarily considered for seawater

applications, 7 percent of the 297 surveyed tubesets used in fresh water were
90-10 copper-nickel. All of the 16 tubesets were at one station. Figure 5-2
shows that 90-10 Cu-Ni has a greater than 90-percent probability of attaining
a 40-year life with 10 percent of the tubes plugged.

Seawater Service. Twenty-six percent of the 226 surveyed tubesets used in sea­
water cooled plants were 90-10 Cu-Ni. There were 58 tubesets. However, the
performance of 90-10 Cu-Ni in seawater was relatively poor. Using the
10-percent of tubes plugged criterion. Figure 5-2 shows that there is only a
40 percent probability of attaining a 40-year life in seawater. Since most of
the 90-10 Cu-Ni problems occur in seawater, the following discussion will be
related primarily to 90-10 Cu-Ni in seawater service.

Erosion-Corrosion Problems. The suggested maximum water velocities for 90-10

Cu-Ni in once-through main condensers are 7.5 feet per second for continuous
service and 8.0 feet per second for peak condition excursions.*

*INCO Bulletin, "The Design and Installation of 90-10 Copper-Nickel Sea Water
Piping Systems," The International Nickel Company, Inc., New York, N.Y.

5-18
The average cooling water velocity in most surface condensers is 7 feet per
second. Some tubes may experience somewhat higher velocities while others
experience lower velocities as determined by the fluid dynamics in a particular
condenser design.

The greatest source of velocity-related problems is partial tube blockage. This

is the result of lodged cleaning tools such as brushes and Amertap balls, or of
lodged marine life such as mussels. Screens did not seem to deter marine life
as small nuclei can pass through the screens, attach themselves to a surface
not subject to chlorination, grow, break loose, and finally become lodged in
the tubes. Lodged obstructions can reportedly cause localized velocities from
two to three times the mean value (142). The effect arises when the presence
of an obstruction reduces the flow through a tube to a fraction of the normal
amount, so that tube friction becomes small and nearly the entire pressure drop
is available to drive water past the obstruction (142).

Other erosion-corrosion problems that occur with 90-10 Cu-Ni and other copper
alloys are related to the release of air bubbles from the cooling water or to
intermittent cavitation. The impingement of rapidly moving water, particularly
where air bubbles are present or where intermittent cavitation occurs, may
result in breakdown of protective films and subsequent severe localized or
generalized attack of the tube surface.* This action commonly occurs near tube
inlet ends where turbulence is highest, but may extend along the entire
length. * ** Severe impingement attack on or near the tube outlet can also
occur when there is excessive downstream vacuum (66).

Sufficient data were not available at the surveyed plants to determine the per­
centage of tubes that failed from erosion-corrosion. However, plant personnel
did indicate the contributing cause of failure of most tubesets. An erosion-
corrosion mechanism was cited as a contributing mechanism in 28 percent of the
corrosion-related citations for 90-10 Cu-Ni in seawater.

*Uhlig, H.H., The Corrosion Handbook, John Wiley & Sons, Inc., New York,
1948, p. 545.
**Fontana, M.B., and N.D. Greene, Corrosion Engineering, McGraw Hill Book Co.,
San Francisco, 1967, P. 78.

5-19
 PERCENT
PLUGGED

40 YEARS

Ul
-20

FRESHWATER

OCEAN WATER

0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
FAILURE RATE

G 1001892-18

Figure 5-2. Failure Rate vs Probability Curves for 90-10 Cu-Ni in Fresh Water
(17 Tubesets) and Seawater Service (55 Tubesets)

i
Effect of Pollution. A primary weakness of 90-10 Cu-Ni and most other copper
alloys is their low resistance to sulfide containing waters. Sulfides generally
result from pollution by organic wastes. Decomposition of these wastes lowers
the dissolved oxygen content of the water and allows anaerobic sulfate reducing
bacteria to convert dissolved sulfate to hydrogen sulfide. Sato (160) has
shown that 0.05 ppm sulfide can markedly increase corrosion on aluminum brass
tubes. Similar effects will occur in 90-10 Cu-Ni, although the pattern of
attack may be more uniform than the other copper-base alloys such as admiralty
brass, aluminum brass, or aluminum bronze (92). Experience (* **) has
demonstrated that copper-nickel alloys are superior to other copper alloys in
tolerating polluted water. INCO has shown, in short-term tests, that 2 to
4 ppm H^S in seawater will increase the corrosion rate of 90-10 Cu-Ni from
3.0 mils per year (in clean seawater) to 34 mils per year. These corrosion
rates were apparently based on weight loss;therefore, the pitting rate may •
have been significantly higher.

Sulfide films form on copper alloys in polluted waters. These films are con­
siderably more noble than the bare metal. Therefore, breaks in the film can
cause severe localized corrosion. This suggests that abrasive cleaning tech­
niques should not be employed in this case.

Only one of the surveyed plants monitored hydrogen sulfide in the cooling
water and, in this case, the sensitivity of the instrumentation was only 1 ppm.
Therefore, no correlation could be made with degree of pollution or with
hydrogen sulfide content. However, the above discussion indicates very low
amounts of disolved hydrogen sulfide (i.e., less than 1 ppm) can cause
significant corrosion. It is therefore possible that many of the pitting
failures are due to sulfide corrosion.

Contributing failure mechanisms were given for many of the tubesets. Some
tubesets had more than one failure mechanism given; others had none. In
order to gain some insight on the extent of a particular mechanism, the
number of times each failure mechanism was cited was tabulated. As shown
in Table 5-8, pitting corrosion was cited 50 percent of the time as a con­
tributing failure mechanism of 90-10 Cu-Ni in seawater.

*Gilbert, P.T., [Link]., July 1959, p. 888.

**Kenworthy, L., Trans. Inst. Mech. Eng., 77, 149, June 1965.

5-21
 Table 5-8

CONTRIBUTING FAILURE MECHANISMS FOR 90-10 Cu-Ni IN SEAWATER

NUMBER OF TIMES
CAUSE CITED PERCENT

General corrosion or specific 15 22

type unknown
Erosion-corrosion 19 28
Pitting 33 50
TOTAL 67 100

Effect of Iron Additions. The presence of ferrous ions in the cooling water
reportedly increases the corrosion resistance of copper-base alloys. The fer­
rous ions, which result from corrosion of upstream components, e.g., inlet
water boxes, or from the deliberate addition of ferrous compounds, e.g.,
ferrous sulfate, cause the formation of an adherent hydrated ferric oxide
(hematite) scale (48). Tests have shown that tubes which are initially covered
with a film of ferric hydroxide are even resistant to polluted seawater (clean
seawater injected with 0.1 ppm sulfide ion) for at least four to five months.
However, the ferric hydroxide deposition is dependent on the initial film already
present; e.g., it will not deposit on tubes already covered with a sulfide
film (80).

An attempt was made to correlate the performance of 90-10 Cu-Ni in seawater with
and without ferrous sulfate additions. Twenty-nine tubesets had been exposed to
water containing ferrous sulfate and 29 tubesets had not. As shown in Figure
5-3, neither system has greater than 50-percent probability of attaining a
40-year life with 10 percent of the tubes plugged. However, the tubesets
exposed to ferrous sulfate do exhibit a better performance.

Insufficient data were available for evaluating ferrous sulfate injection

methods. However, continuous injection is reportedly far superior to inter­
mittent injections of the same gross amount (85). A minimum of 0.03 ppm con­
tinuous injection is needed apparently for protection (160).

70-30 Copper-Nickel, CN, CA715

Freshwater Service. Only six of the 297 surveyed tubesets used in freshwater
service were 70-30 Cu-Ni. Failure rate data were not available for these tube-
sets, therefore no performance evaluations were possible for freshwater service.

5-22
 PERCENT
PLUGGED

1001
OPERATING TIME, HR x

40 YEARS
%
ROBABILITY,

FeSO, ADDED

NO FeS04
ADDITION

10 --------------- -------------—rZLZ------ ------- ------ -----———------ ----------------- ------------- --------- ------- -----------L.'1 ............................ .................L--------------- -----------—L—LJ
0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
FAILURE RATE

Figure 5-3. Failure Rate vs Probability Curves for 90-10 Cu-Ni in Seawater Service,
With (29 Tubesets), and Without (29 Tubesets) FeSO^ Additions.
Cooling Tower Service. Twelve of the surveyed tubesets used in cooling tower
service were 70-30 Cu-Ni. All but one exhibited a zero failure rate.

Seawater Service. Twenty-five percent of the 226 tubesets surveyed at seawater-

cooled plants were 70-30 Cu-Ni. Of the 51 tubesets for which failure rate data
were available, 45 tubesets or 88 percent were in air removal sections. This
indicates that 70-30 Cu-Ni is used primarily for the air removal sections of
seawater-cooled condensers. Figure 5-4 shows that 70-30 Cu-Ni is not well
suited for seawater service if high reliability and long service life are
required. The 6 main-body tubesets had a median FR of 0.102, whereas the 45 air
removal tubesets had a median FR of 0.060. Using the 10-percent tubes plugged
criterion, 70-30 Cu-Ni is shown to have less than 70 percent probability of
providing a 40-year service life. This is a better performance rating than the
40-percent probability for 90-10 Cu-Ni under the same conditions.

Erosion-Corrosion Problems. The recommended maximum design water velocity of

70-30 Cu-Ni tubes in seawater-cooled condensers is 12 feet per second. This is
higher than that for 90-10 Cu-Ni.* However, as discussed in the 90-10
Cu-Ni subsection, much higher velocities can result from partial tube blockages.
Referring to Table 5-9, 44 percent of the citations for corrosion-related
failures showed erosion-corrosion to be a contributing cause.

Table 5-9

CONTRIBUTING CORROSION FAILURE MECHANISMS FOR

70-30 Cu-Ni IN SEAWATER

NUMBER OF TIMES
CAUSE CITED PERCENT

General corrosion or specific 15 22

type unknown
Erosion-corrosion 32 44
Pitting 26 36
TOTAL 73 100

*La Que, F.L., Marine Corrosion, John Wiley & Sons, New York, 1975.

5-24
 1000

PERCENT
PLUGGED

1001
OPERATING TIME, HR x
40 YEARS
%
'ROBABILITY,

0.2
FAILURE RATE

Figure 5-4. Failure Rate vs Probability Curve for 70-30 Cu-Ni in Seawater (51 Tubesets,
of Which 6 are Main Body and 45 are Air Removal).
Effect of Pollution. In waters containing sulfide ions, 70-30 Cu-Ni is highly
susceptible to corrosion. Since only one plant monitored hydrogen sulfide
concentration, no correlations were possible. The cooling water at one plant
which had just retubed with 70-30 Cu-Ni became contaminated with an unknown
concentration of hydrogen sulfide. These tubesets lasted less than one year.
Another station, which had enjoyed relatively good service with 70-30 Cu-Ni for
several years, began to experience numerous failures when the intake canal was
stirred up during construction of the intake system for a new unit. Hydrogen
sulfide was suspected of causing the failures although H^S was not detected by
the monitoring system. However, the sensitivity of the instrumentation was
only 1.0 ppm H^S. As stated in the 90-10 Cu-Ni subsection, sulfide concentra­
tions of 0.05 ppm have been shown to significantly increase corrosion.

Effect of Iron Additions. Twenty-four 70-30 Cu-Ni tubesets had been exposed
to water containing ferrous sulfate and 33 had not. Figure 5-5 indicates that
a significant increase in performance was achieved when ferrous sulfate was
added to the water. Greater than 90 percent probability of achieving a 40-year
life with only 2 percent of the tubes plugged is indicated when ferrous sulfate
was added. However, these data should be considered cautiously. The tubesets
in the two stations which were known to have sulfide-contaminated water did not
use ferrous sulfate additions.

Aluminum Brass, BA, CA687

Freshwater and Cooling Tower Service. None of the stations surveyed used alu­
minum brass in fresh water or cooling tower waters.

Seawater Service. Excessive impingement-attack failures of admiralty brass

tubes in turbine condensers used in the British navy prompted the research
programs that lead to the development of aluminum brass, as well as the copper-
nickel alloys. The goal was to provide greater reliability in turbine condenser
applications on ships, seashore power plants, and other heat exchangers using
seawater for cooling.* Twenty-three percent of the 226 surveyed tubesets used
in seawater-cooled plants were aluminum brass. Figure 5-6 shows that there is
less than 50 percent probability that aluminum brass tubes will provide a 40-
year life in seawater with not more than 10 percent tubes plugged. Therefore,
this alloy appears to be unsatisfactory for service that requires high reli­
ability and a long service life.

*La Que, F.L., op cit.

5-26
 PERCENT
PLUGGED

1001
OPERATING TIME, HR x

40 YEARS

FeSO. ADDED
%
'ROBABILITY,

NO FeSO ADDED

0,2
FAILURE RATE

Figure 5-5. Failure Rate vs Probability Curves for 70-30 Cu-Ni in Seawater With
(24 Tubesets) and Without (33 Tubesets) FeSO^ Additions.
 PERCENT
PLUGGED

I 400

40 YEARS
H 300

a: 200

U1
ss-

10 --- -—---- J------ ---- --- —*1—1—1——----------- ------ ---- --- ----- ——-I——--------- ------ ---- 1——1---——1-1-1
0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
FAILURE RATE

Figure 5-6 Failure Rate vs Probability Curve for Aluminum Brass in Seawater
(37 Tubesets).
Erosion-Corrosion Problem. The recommended maximum-design water velocity for
aluminum brass tubes in seawater-cooled condensers is 8 feet per second.*
Therefore, small deviations from the normal condenser design velocity, i.e.,
7 feet per second, will cause problems in aluminum brass tubing, even if
partial tube blockages and localized turbulence can be eliminated. This is
reinforced by the data in Table 5-10 which show that erosion-corrosion was
cited 65 percent of the time as contributing to corrosion related failures
of aluminum brass in seawater.

Table 5-10

CONTRIBUTING CORROSION FAILURE MECHANISMS FOR

ALUMINUM BRASS IN SEAWATER

NUMBER OF TIMES
CAUSE CITED PERCENT

General corrosion or specific 6 22

type unknown
Erosion-corrosion 26 65
Pitting 8 20
TOTAL 40 100

Effect of Pollution. Aluminum brass is highly susceptible to waters containing

sulfide ions. INCO test data indicate that aluminum brass has a susceptibility
similar to that of the copper-nickels to polluted seawater.**

Effect of Iron Additions. Thirty-two aluminum brass tubesets had been exposed
to water with no ferrous sulfate additions while only four tubesets had been
exposed to water with ferrous sulfate. These four tubesets were located at
one station and were exposed to one of the most polluted bodies of water in
the country. Figure 5-7 indicates that in this case, poorer performance was
attained with the ferrous sulfate additions than with tubesets not exposed to
ferrous sulfate. However, as discussed earlier, the ferric hydroxide deposi­
tion is dependent on the initial film already present and it will not deposit
on tubes already covered with a sulfide film. Therefore, ferrous sulfate
offers no protection for tubes already covered with a sulfide film. The
operating history of the particular condenser was such that sulfide-free
surfaces were probably never available. Therefore, the curves in Figure 5-7
may be misleading.

*La Que, F.L., op cit .

^Correspondence from Development and Research Department, International
Nickel Company, Inc.

5-29
 'PERCENT
PLUGGED

40 YEARS
NOTE: SEE TEXT FOR
QUALIFICATIONS
ON THIS FIGURE.

Ol
-30

ADDE
NO FeSO ADDED

0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
FAILURE RATE

Figure 5-7 Failure Rate vs Probability Curves for Aluminum Brass in Seawater
Service With (4 Tubesets) and Without (32 Tubesets) FeSO^ Additions.
Aluminum Bronze, AB, CA608

Although aluminum bronzes are commonly used in seawater, very few applications
involve condenser tubes. Only 12 tubesets, all in seawater-cooled plants,
utilized aluminum bronze tubes. These were installed at two stations. Eight
tubesets have been replaced with another alloy and the other 4 soon will be.
Figure 5-8 indicates a low probability that aluminum bronze will provide a
40-year service life with less than 10 percent of the tubes plugged.

Because of the small data base, no evaluation could be made of the effects of
ferrous sulfate addition. Ferrous sulfate was used at one station and not at
the other. High failure rates were observed at both stations.

Contributing failure mechanism citations are given in Table 5-11. Most failures
appear to be related to waterside corrosion; pitting corrosion is the mechanism
cited most often.

Table 5-11

CONTRIBUTING FAILURE MECHANISMS FOR ALUMINUM BRONZE

NUMBER OF TIMES
CAUSE CITED PERCENT

General corrosion or specific 6 25

type unknown
Erosion-corrosion 6 25
Pitting 12 50
TOTAL 24 100

STAINLESS STEELS

Since the early 1960s, increasing interest has been shown in stainless steel
tubing for large surface condensers. In this survey, 169 or 29 percent of the
580 total tubesets were stainless steel as indicated in Table 5-1. Likewise,
Table 5-4 shows that 20 percent of the tubes and 20 percent of the total con­
densing surface area in the surveyed plants were fabricated from either Type
304 or Type 316 stainless steel. Type 304 accounts for 99.6 percent of the
stainless steel condenser tubing surface area. Table 5-12 lists the composi­
tion of these stainless steels and the designations used in this study.

5-31
 1000

800

600
PERCENT
o PLUGGED
o
x 2 X5 Xi o
£C 400
X
UJ 40 YEARS
I 300

C3 |N
Z
|
< 200
ir 1
m
O
X
1
1
100 1

1
80
1
in
1
-32

60
1

1
1
40
m
< 1
CD
O
DC
1

1
1
20 1

[
1

1
10
0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
FAILURE RATE

Figure 5-8. Failure Rate vs Probability Curve for Aluminum Bronze in Seawater Service
(12 Tubesets).
 Table 5-12

STAINLESS STEEL -
CHEMICAL COMPOSITION AND DESIGNATIONS

COMPOSITION
Study Manga­ Phos­
Desig­ Carbon, nese , phorus , Sulfur, Silicon, Chro­ Molyb­
SS Alloy nation max max max max max Nickel mium denum

Type 304 SS 0.08 2.00 0.040 0.030 0.75 8.0 18.0 ___

to to
11.0 20.0

Type 316 ST 0.08 2.00 0.040 0.030 0.75 11.0 16.0 2.00
to to to
14.0 18.0 3.00

Erosion-Corrosion. Stainless steels show much greater resistance to erosion-

corrosion than copper-base alloys. There is no recommended maximum water
velocity design limit for stainless steel tubing in condenser service. Erosion
by cooling waters has not been a problem. In fact, erosive waters such as those
containing large amounts of suspended solids actually improve the performance of
stainless tubing by maintaining a cleaner surface. In many cases, stainless
steel tubing has been used for the steam impingement areas in order to take
advantage of its erosion resistance where a distinct impingement zone could be
determined. Stainless steel was used in these sections in 76 percent of the
surveyed freshwater condensers.

Ammonia Attack. Stainless steels are immune to ammonia attack and to ammonia-
induced stress corrosion cracking. For this reason, they are also used in air
removal sections where the ammonia concentration may be high. In the surveyed
freshwater condensers, stainless steels were used in 54 percent of the identi­
fiable air removal sections.

Chloride Stress Corrosion Cracking. On the basis of both the field survey and
the literature search, as well as discussions with cognizant individuals,
chloride-induced stress corrosion cracking has not been observed in stainless
steel tubing used in power plant condensers. The cooling water temperature in
most condensers is below the so-called critical temperature for stress corrosion
cracking, i.e., 120 to 140°F in practical applications. Stress corrosion

5-33
cracking perhaps could occur when cooling towers are used which result in
higher temperatures (100 to 120°F). However, it is more likely that stainless
steel will fail first by other mechanisms, e.g., pitting corrosion.

Pitting Corrosion. One weakness of Type 304 and 316 stainless steels is their
susceptibility to pitting corrosion. This susceptibility is increased when the
cooling water has a high chloride concentration, e.g., ocean water or some cool­
ing tower waters. Pitting attack is aggravated by stagnant or low flow condi­
tions in which solids can settle on the tube surface. Small concentrations of
manganese were identified as the cause of pitting corrosion in at least one case.

The addition of molybdenum to stainless steels increases their resistance to

pitting corrosion. Type 316 has a molybdenum content of 2 to 3 percent. This
amount of molybdenum apparently is not required in most condensers cooled by
fresh water with a low chloride concentration, but it is too low for ocean and
brackish cooling waters which have a high chloride concentration.

Allegheny Ludlum alloy 6X, which has a nominal molybdenum content of 6.5, shows
promise in overcoming the pitting problem in high chloride waters. Test tubes
have given up to six years of service in seawater cooled condensers without a
failure.* Recently, two seawater-cooled condensers have been completely retubed
with AL6X.

Type 304, Stainless Steel, SS

Freshwater Service. Thirty-nine percent of the 297 surveyed tubesets in fresh

water were Type 304 stainless steel. For the 95 tubesets with known failure
rates, 27 percent were used in the main-body section, 45 percent were used in
the air removal section, and 28 percent were used in impingement sections.

Figure 5-9 shows that when only main body and air removal tubesets are con­
sidered (Curves A and B), the performance of Type 304 stainless steel in fresh
water is excellent. A greater than 90 percent probability is indicated that the
tubes will provide a 40-year life with less than 2 percent of the tubes plugged.
Curve B shows the probability versus failure rate obtained when all available
total failure rate data for Type 304 stainless steel tubes in freshwater service

*Personal Communication from J. Maurer of Allegheny Ludlum on November 15, 1976.

5-34
OPERATING TIME, HR x 1Q0I
PERCENT
PLUGGED

40 YEARS
%
ROBABILITY,

0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
FAILURE RATE

Figure 5-9. Failure Rate vs Probability Curves for Type 304 Stainless Steel in
Freshwater Service; Curve A-A and M Tubesets (Guard Plugs at One
Station Deleted) (64 Tubesets); Curve B-A and M Tubesets (No Exclu­
sions) ; Curve C-I Tubesets (27 Tubesets).
is plotted. An unusually large number of guard plugs were used in one condenser
although only two leaking tubes had actually been discovered. Since this large
number of guard plugs distorts the data. Curve A was plotted to more realisti­
cally show the performance of SS 304 in main body and air removal service.

The performance of impingement sections (Curve C) is somewhat lower. Using the

10-percent tubes plugged criterion, less than 60 percent probability is
indicated that these tubes will provide a 40-year service life. Failures in the
impingement section are usually due to projectiles or steam impingement.
Failures from such causes are not a function of the tubing material. Such
failures are also independent of time. Accordingly, calculation of the failure
rate for impingement sectors is of questionable value.

Cooling Tower Service. Thirty-seven percent of the tubesets in cooling tower

service were Type 304 stainless steel. Although Figure 5-10 indicates that
Type 304 stainless steel will provide good performance in cooling tower service,
the data may be misleading. Failure rate data were available for only 13 of the
21 tubesets surveyed. Severe corrosion occurred in the tubesets for which no
failure rate data were available. Therefore, the performance curve in Figure
5-10 would probably be much lower if data from these tubesets were available or
if the data base were larger.

Contributing Failure Mechanisms. As discussed earlier, contributing failure

mechanisms were given for many of the tubesets. Some tubesets had more than
one failure mechanism given. Table 5-13 shows the distribution of all contribu­
ting failure mechanisms (similar tables in the copper alloy section listed only
corrosion related failures) for Type 304 stainless steel in fresh water.

The significant feature of Table 5-13 is that only 2 percent of the failure
mechanism citations were related to water-side corrosion. Eighty-nine percent
of these failures were related to steam-side mechanical failures and 9 percent
to faulty tube rolling.

Insufficient data were available for comparing contributing failure mechanisms

in the cooling tower tubesets. As noted above, at one plant with a cooling
tower Type 304 tubes experienced severe pitting corrosion.

5-36
 PERCENT
PLUGGED

40 YEARS
5 300

0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
FAILURE RATE
Figure 5-10. Failure Rate vs Probability Curve for Type 304 Stainless Steel
in Cooling Tower Service (13 Tubesets). (Note: As explained
in the text, this curve may be misleading.)
 Table 5-13

CONTRIBUTING FAILURE MECHANISMS

FOR TYPE 304 STAINLESS STEEL
IN FRESH WATER

NUMBER OF TIMES
CAUSE CITED PERCENT
Falling objects 36 38
Mechanical 17 18
Steam erosion 16 17
Vibration 15 16
Faulty tube rolling 8 9
General corrosion or specific 2 2
type unknown

TOTAL 94 100

Effect of Chlorides. It is well known that chloride ions increase the suscepti­
bility of stainless steel to pitting and crevice corrosion. It has also been
shown that susceptibility to pitting and crevice corrosion increases with
temperature.* **

However, the critical chloride concentrations and temperatures above which

pitting in Type 304 and other stainless steels will occur have not been defined.
Chloride concentrations in the cooling water ranged from 7 to approximately
120 ppm at the 12 surveyed plants which used Type 304 stainless steel tubing.
The average and median failure rates for the 21 tubesets (2 stations) exposed
to approximately 100 ppm chloride cooling water were 0.074 and 0.0, respectively.
These values are considerably less than the 0.333 failure rate which corresponds
to a 40-year life with less than 10 percent of the tubes plugged. The average
to date service life of these tubesets is 9.5 years and all are original tubesets.
Station practices did not include special cleaning or maintenance procedures
for these tubes. Therefore, it would appear that cooling water chloride concen­
trations of approximately 100 ppm may not be a problem for Type 304 stainless
steel tubing when the condenser operating temperatures are approximately 100 to
110°F.

*Brigham, R.J. and E.W. Tozer, "Temperature as a Pitting Criteria," Corrosion

Vol 29, No. 1, January 1973, p 33.
^Correspondence from Development and Research Department, International Nickel
Company, Inc.

5-38
Maintenance and Cleaning Procedures. Table 5-14 summarizes the layup and
cleaning procedures used by those utilities that had Type 304 stainless steel
condenser tubing in freshwater service.

Type 316 Stainless Steel, ST

Freshwater Service. Only 3 percent of the surveyed freshwater tubesets were

Type 316 stainless steel. These were all used in either the air removal or
impingement sections of the condensers. The median failure rate for these tube-
sets was 0.00. Contributing failure mechanism data were not available for Type
316 stainless steel tubesets used in fresh water.

Cooling Tower Service. Type 316 stainless steel condenser tubes were not used
at any of the surveyed stations which had cooling towers.

Seawater Service. Ten percent of the 226 tubesets in seawater-cooled condensers

were Type 316 stainless steel. The failure rates (data were available for only
9 of these tubesets) ranged from 0.116 to 28.571. The average and median failure
rates were 5.568 and 1.933, respectively. Such failure rates confirm that
Type 316 stainless steel is not well suited for seawater or brackish water
service. Specific failure mechanisms were cited 27 times for Type 316 stain­
less steel in seawater. Pitting corrosion accounted for 59 percent of the
citations, erosion-corrosion for 30 percent, and general corrosion or specific-
type-unknown for 11 percent.

TITANIUM, TI

Titanium as a tubing material for domestic steam power plant surface condensers
first was investigated in 1959 when 7 test tubes were installed (262).* The
first condenser with all-titanium tubes was put into service in May 1972.**

*"Titanium Tubing for Surface Condenser Heat Exchanger Service, Bulletin

SC-3," Titanium Metals Corporation of America, April 1975, p.5.
**"Installation List, TIMET Codeweld Condenser Tube," Titanium Metals
Corporation of America, April 1976.

5-39
 Table 5-14

MAINTENANCE AND CLEANING PROCEDURES USED

FOR TYPE 304 STAINLESS STEEL TUBING

COOLING WATER
CHLORIDE CONCEN­ NORMAL CLEANING LAYUP
STATION TRATIONS, ppm PROCEDURES DURING OUTAGES

A 7-14 Backwash once a week or Drain and let

as needed. Cleaning with dry
plugs once a year

B 14-72 Water lance twice a year Drain only

C ^23 None Unknown

D 100-120 None Drain and dry

E 7-10 Short-Leave
circulating
water pumps
on Long-Drain,
dry, and clean
with nylon
brushes

F Not available Amertap Short-None

(Conductivity: Long-Drain only
60-90umho)

G 10-13 Backwash Leave circu­

lating water
pumps on

H 100-120 None Drain and dry

I Not available None Leave circu­

lating water
pumps on

J Not available None None

(VL35TDS)

K 12-13 Backwash every shift; Leave circu­

use water and air gun lating water
twice a year pumps on
If entry re­
quired, back­
wash, drain,
and flush with
deionized water

L 60 Backwash when needed; Drain only

clean with brushes
when needed

5-40
Major installations of titanium tubing have been in service in seawater distil­
lation plants since 1965 (242). With regard to steam condensers using titanium
tubing, the following describes a typical operating installation. This descrip­
tion is not intended to stereotype the application of titanium but to broadly
represent its present use in operating plants as of the time of this survey.

• Tubing — manufactured to ASTM specification B-338, Grade 2;

7/8 in. (2.22 cm) or 1 in. (2.54 cm) O.D., 22 BWG wall thick­
ness; seam welded; supplied by Titanium Metals Corporation of
America (TIMET)
• Tube-to-Tubesheet Joint — rolled, often with grooved tube-
sheet holes; muntz metal (or aluminum bronze) tubesheet
material (most tubesheets were existing in condensers retubed
with titanium)
® Circulating Water — high in dissolved solids; often low in
dissolved oxygen and with high pollutant concentrations
• Installed by retubing existing condenser; installation in new
condensers designed for titanium tubes has occurred only
recently

Under the general conditions outlined above, the performance of the material
has been very good.

The basic material specification requirements for titanium as used in conden­

sers* (6-p.124) are given by ASTM B338-76, Grade 2. The ASTM chemical composi­
tion limits are as follows:

Nitrogen, max 0.03%

Carbon, max 0.10%
Hydrogen, max 0.015%
Iron, max 0.30%
Oxygen, max 0.25%
Other elements (each), max 0.1%
Other elements (total), max 0.4%
Titanium (by difference) remainder

*"Titanium Tubing, Bulletin SC-3," p.7.

5-41
Commercially pure titanium is unique in that it is the only condenser tube
material that is used in an unalloyed form.

Seamless titanium has been tested by some utilities in the U.S., but it is not
cost competitive with seam-welded tubing. The Japanese use seam-welded tubing
in their condenser applications (135). In Britain seamless titanium tubing has
been used in power station condensers (6-pp. 124, 127-128).

TIMET, which has supplied the titanium tubing for all U.S. condenser installa­
tions except one since Septemeber 1976, has provided eddy current, pneumatic,
and ultrasonic inspections on all production tubing. Hydrostatic testing is
provided at extra cost when requested.

Survey Results* •

One of the original objectives of the study was to survey steam power plants
containing a significant number of titanium condenser tubes. Completed mail
survey forms were received from all 14 units that had titanium condenser tubes
in service. Seven of the responses stated that zero failures of titanium
tubing had occurred. The remaining 7 responses were understood as indicating
very few, if any, failures, but could not be definitely interpreted as zero
failures. The following points will gnerally summarize the mail survey infor­
mation received:

• Ten of the 14 units were retubed in whole or in part with

titanium. The previous tubing material had been aluminum
brass, aluminum bronze, 90-10 copper-nickel, or 70-30 copper-
nickel.
• The average dissolved solids in the once-through circulating
water ranged from 5,000 to 32,000 ppm at the 10 retubed units.
• Two of the units with original titanium tubing use fresh to
slightly brackish cooling water (including one cooling tower
installation). At the other 2 locations, seawater is used
for cooling water.
« The Amertap cleaning system is used continuously or inter­
mittently on 3 of the 14 installations (it is available but
not used on a fourth unit); 2 other units clean intermit­
tently with brushes or plugs; 10 units reported the use of
chlorination.

5-42
Of these 14 operating units with titanium tubing, 10 were visited during the
field survey. Also included in the field survey were two units that had
titanium tubing installed subsequent to the mail survey. The extent of the
survey and performance results are summarized below:

• Eleven of the 12 units were not originally tubed with titanium.

• Thirty-one tubesets of titanium were defined at the 12 units
included in the field survey.
• A total of 164,000 titanium tubes was included. The average
operating life was about 14,000 hours.
• The seven titanium tubes initially installed in 1959 for test
purposes in polluted, brackish cooling water have performed
without evidence of corrosion after approximately 91,000 hours
of service as of April 1973 (127,262).
• The data for all known titanium failures as of early December
1976 show 22 tubesets with zero failures and 9 tubesets with
known failures. For the 9 tubesets with failures, the sta­
tions reported the following:

Cause of Tube Failures Number of Tubesets Number of Failed Tubes

Steam Impingement 1* 15
Vibration Damage 5** 26
Mechanical Damage 3* 6
Unknown 1 1
TOTAL 9 48

*0ne tubeset had 2 types of failure causes.

**At one unit, approximately 300 plugged tubes were replaced recently with new
titanium tubes. The new tubes were staked to control vibration.

The overall performance of titanium is indicated in Figure 5-11. The field

survey data show that titanium condenser tubes have a high reliability.
Titanium tubes have a 87 percent probability of operating for 40 years with not
more than 2 percent plugged tubes and a 94 percent probability of 40-years
service with not more than 1 percent plugged tubes. The median failure rate

5-43
 HR)

%>
ItT
OPERATING TIME (IN
’ROBABILITY, %

I I I I I I
0.03 0.04 0.05 0.06 0.08 0.1 0.4 0.5 0.6 0.8 TO
FAILURE RATE

Figure 5-11. Failure Rate vs Operating Hours and Failure Rate

vs Probability for Titanium Tubing

5-44
value for the tubsets surveyed is zero. Some qualifications should be observed
when considering these probabilities and statistics:

• The known causes of failure have been vibration, steam impinge­

ment, or mechanical damage. Vibration damage generally tends
to manifest itself early in the life of a tubeset. Steam
impingement also usually occurs early in the service life but
can result from additions of drain lines, etc, later in the life
of the condenser. These two problems can usually be permanently
solved by staking and baffling of steam jets, respectively.
Mechanical damage (turbine-generated missiles, dropped tools,
broken baffles) can occur at any time. It is independent of
tube material and cannot be prevented from recurring although
precautions can be taken to limit its frequency. Given such
failure mechanisms, the above anticipated service life proba­
bilities are probably low.
• The average tube age is about 14,000 operating hours (about 2
calendar years). The age of the oldest titanium installation
surveyed is 5-1/2 years. Although titanium, unlike other tubing
materials in condenser service, is virtually immune to pitting,
corrosion, erosion, stress corrosion, etc (as described later) any
possible long-term deleterious factors could not be determined due
to the relatively short lives of the surveyed tubesets. The
above service life probabilities should be viewed with this in
mind.
• The above service life probabilities include the effect of plugs
in tubes that have not failed (guard plugs and plugs resulting from
removal of inspection tubes). If only failed tubes were included,
the above probabilities would be substantially higher.

Discussion Items

A number of techanical factors relate to the use and performance of titanium

tubing in surface condensers. These are enumerated below, followed by a
synopsis of the information found in the literature and field investigation.

# Vibration
# Tube-to-Tubesheet Joints
0 Heat Transfer/Fouling
® Hydriding/Galvanic Corrosion
# Other Corrosion and Erosion Factors

5-45
Vibration. Approximately one-half of the relatively few failures of titanium
tubing are directly attributable to vibration damage.

The damage itself is manifested by: 1) fatigue failure of the tube where it
passes through a support plate; or 2) flattened regions on the outside surface
of the tube along its length due to collision with adjacent tubes which can
cause eventual rupture.

This failure mechanism is by no means peculiar to titanium. The field survey

found instances of vibration damage in admiralty, aluminum brass, and stainless
steel tubing. Usually, tube vibration is caused by high shell-side cross-flow
steam velocities at the top or upper sides of the tubing bundles. It also occurs
adjacent to poorly baffled auxiliary dump and drain lines. An instance of tube
damage by vibration in an air removal section was found during the field survey.
Improper support plate spacing or steam flow distribution, or unusual condenser
operating conditions are typical sources of the problem.

(NOTE: The very brief discussion of vibration that follows is intended to show
only qualitative relationships. The reader is directed to the indicated
references for a more complete treatment.)

The major excitation force causing condenser tube vibration results from the
cross-flow steam velocity. The resulting drag force causes tubes to vibrate
at their natural frequency given by the equation (235-p.l65; also see 4 and 354):
1/2
f
n 4
WL

Where: f = natural frequency

g = gravitation constant
E = modulus of elasticity
I = moment of inertia of tube cross-section
W = weight per unit length of tube filled with water
L = tube span between supports
C = end support constant

5-46
The deflection of the vibrating tube is given proportionally as
4
V
Y ^ k
El

Where:
Y = deflection or amplitude
= (drag) force per unit length causing the
deflection
k = end support constant
Also affecting the amplitude of vibration is the amount of system damping.
To avoid impact with adjacent tubes at their midspan, the tube vibration
amplitude must be restricted. For a given steam flow condition (W ), amplitude
reduction may be accomplished by: 1) decreasing tube span (L); 2) increasing
the modulus of elasticity (E); or 3) increasing the moment of intertia (I). The
natural frequency will simultaneously increase with any of these changes as well,
thus indicating an increase in tube stiffness.

Comparing titanium to the copper-base and stainless steel tubing alloys under
identical conditions it is more susceptible to this type of damage because:
1) titanium has a relatively low modulus of elasticity; and 2) thinner-walled
titanium tubing has a lower moment of inertia. Both of these result in a tube
with higher vibration amplitudes which can be reduced by decreasing the tube
span (117).

Correction for this potential problem, as it relates to titanium-tubed condensers,

depends on the particular situation. When designing a new condenser for use
with titanium tubes, suitably close tube support plate spacing can be pro­
vided (133). When retubing an existing condenser, a careful evaluation of the
potential for vibration should be made (132). In this regard, one utility
decided not to retube one of its units with titanium because of a previous
history of vibration damage to copper-base alloy tubing. In a retubing situa­
tion where the possibility of vibration exists, the tubing can be staked by any
of a number of methods thereby reducing the effective tube span. Finally, in
the instance of locally high steam velocities, staking of the tubing or disper­
sion of the jetting steam can control or eliminate the problem.

5-47
Table 5-15 shows the reported data from the field survey relating to vibration
in titanium tubes. The units are listed in order of increasing tube support
spacing. This table shows no apparent relationship between support spacing and
vibration damage. Other factors such as steam flow velocities and distribution,
peculiar to each unit, also play an important role.

Table 5-15

FIELD SURVEY DATA

TITANIUM AND VIBRATION

APPROXIMATE REPORTED TUBE TUBE

TUBE SUPPORT VIBRATION O.D. , GAGE
CONDENSER* SPACING, in. (cm) TUBING STAKED FAILURES in. (cm) BWG
A 30 ( 76) Yes** - through­ 1 tube 7/8 (2.22) 22
out all showed
bundles No - minor
wear
B 36 ( 91) Yes** - at top some 7/8 (2.22) 22
of on
bundles Yes- side of
bundles
C 38-1/2 ( 98) No No 7/8 (2.22) 22
D 38-1/2 ( 98) No No 7/8 (2.22) 22
E 38-1/2 ( 98) No No 7/8 (2.22) 22
F 41 (104) NO No 1 (2.54) 22
G 42 (107) No No 1 (2.54) 20
H 42 (107) No No 1 (2.54) 22
I 42 (107) No - but will be Yes - some 1 (2.54) 22
done at
top
J 51 (130) No No 1 (2.54) 22
K 57 (145) Yes** - at top Yes - numer­ 1 (2.54) 22
of ous on
bundles side of
bundles

*One of the 12 units surveyed has no titanium impingement tubes.

**Staking done at time of retubing with titanium.

5-48
Tube-to-Tubesheet Joints. In addition to the failure of the tube itself, the
other potential major source of circulating water inleakage into the condenser
is the tube-to-tubesheet joint. In power industry condensers, this joint is
typically made by expanding the tube into the tubesheet — forming an inter­
ference fit. Some condensers have welded or packed joints. Utility personnel
interviewed during the field survey reported the incidence of detectable leakage
through tube-to-tubesheet joints is very low. Either joint integrity is suf­
ficiently good to prevent leakage or the amount of leakage is too small to be
revealed by the leak detection and location methods used. For the titanium
tubing included in this survey, all tube-to-tubesheet joints were made by roller
expanding the tubes into single tubesheets at each end of the tubes. The British
use roller expanded tubes but also have double tubesheets in some of their
titanium-tubed condensers. The tubes are rolled into both tubesheets at each
end (6-p.l26).

Welded joints have been little used in surface condensers with titanium tubing.
Some desalination (242) , chemical, and refinery applications have used welded
joints with titanium tubes and tubesheets. The Japanese have used such "all­
titanium" installations in industrial turbine condensers (133). The British
have developed a method of explosion-welding a seamless titanium tube into a
tubesheet of dissimilar metal (336; 339; 235-p.81). Of principal concern when
fusion-welding titanium is the maintenance of a complete inert gas atmosphere
to prevent severe embrittlement from atmospheric contaminants. To date, fabri­
cating shops have best been able to provide the required craftsmanship, equip­
ment, and tight control of welding conditions. But present-day condensers have
required field erection because of their large size. Modular construction of
condensers using shop-welded joints of lighter weight titanium tubing may pro­
vide a solution to this dilemma (235-p.87; 354). Previous experience with
welded joints in chemical processing plants has shown them to be strong and
corrosion-free (262).

5-49
As with welded joints, joint integrity, joint strength, and material compatibility
are important factors with rolled titanium joints. The integrity of a rolled
joint is normally tested following installation by flooding the steam side of
the condenser. Any joint found leaking is normally rerolled to seal the joint.
Initial joint leakage data were obtained for 3 units in which titanium tubes
were installed. Discounting tubes that were inadvertently bypassed during the
rolling operation, 1 unit containing about 15,000 tubes had about 35 tubes that
required rerolling. Another required rerolling of about 20 of 14,000 tubes (132).
The third required rerolling of about 110 of 13,000 tubes. In the third instal­
lation, half of the tube joints had been rolled by an inexperienced boilermaker;
some of the tubes had been trimmed by a method that resulted in a loosening of
joints. The first 2 units have grooved, aluminum bronze tubesheets; the third
unit has ungrooved muntz metal tubesheets. Because of the tendency of titanium
to work-harden, rerolling of tubes must be done with care. Comparative data for
other tube materials were not obtained.

The mechanical strength of the joint is measured by the amount of force required
to pull the joint apart. Numerous factors affect this pullout strength including
tube material, tubesheet material, size of contact area, whether the hole is
grooved, original material dimensions, amount of interference fit, type of tube
expander used, and workmanship. Various utilities, condenser manufacturers,
tubing fabricators, and tubing installers have tested the pullout strengths of
differing materials and geometries. Although no compilation of information was
made, some general observations are offered:

• Grooves in the tubesheet holes will improve joint strength. A

number of the titanium installations have no tubesheet grooves
and are operating satisfactorily.
• Higher yield strength aluminum-bronze tubesheets provide a
stronger joint than does muntz metal. About two-thirds of
the surveyed units with titanium tubing had muntz metal
tubesheets.

The apparent trend in new condensers designed for titanium tubing is to use
aluminum-bronze or titanium-clad steel tubesheets.

5-50
A flare on the inlet end of the condenser tubes historically has been provided
to reduce power-wasting turbulence and to alleviate erosion of the tubing. Since
titanium has very good resistance to erosion (as described later), flaring of
the tubing is not necessary to reduce erosion. To improve hydraulic performance,
titanium tubes can be flared if necessary precautions are taken to avoid loosen­
ing the rolled joint or splitting the tube end. All of the field survey units
except two had flared tube inlets.

Heat Transfer/Fouling. When comparing the performance of titanium to the copper-

base alloys in particular, heat transfer and fouling characteristics are fre­
quently mentioned. The copper-base alloys have a higher thermal conductivity
than titanium and a natural, surface poisoning effect on biofouling organisms.
Various design features are being used or proposed to enhance the thermal per­
formance of titanium (6-pp. 127, 125; 133), as follows:

• Thinner-walled tubing, 25 BWG instead of the more typical

22 BWG, would further reduce heat transfer resistance.
• Increased tube water velocity from the normal 7-8 fps (2.2
meters per second) to 10 fps (3.0 meters per second) or higher
to decrease the fouling and silting of the tubes (133; 134;
6-p.l27), improve the water-side heat transfer coefficient,
and, consequently, reduce the required surface area of the
condenser and decrease circulating water temperature rise.
Such benefits must be balanced against the additional pumping
costs that result. Other factors such as water-box design and
flow distribution also require careful evaluation.

The thermal performance of titanium appears to be good. Current design

practices employ cleanliness ratios for titanium in the range of 85 to 100
percent — similar to those for stainless steel tubing. The British found that
although 70-30 copper-nickel tubes had an initially higher overall heat transfer
coefficient than titanium, after 14 months of operation the titanium performance
was better than that of copper-nickel due to lower fouling (6-pp. 127;129). The
Japanese found that after 8 months of operation without tube cleaning, titanium
tubing in a small, industrial, steam condenser was still performing at the
original 90 percent design cleanliness (133) . One installation in this country
is experiencing problems in maintaining a design cleanliness of 85 percent in
spite of frequent cleaning of the tubes with plugs. The fouling condition is
apparently caused by a high silt concentration in the water, aggravated by the

5-51
lack of a circulating water biocide. Although quantitative data from the field
surveyed units are lacking, stations retubed with titanium are experiencing
comparable or slightly better thermal performance than with the previous copper-
based alloy tubing.

The two principal sources of condenser tube fouling are the inorganic deposits
and the deposits associated with marine and freshwater organisms. Broadly
speaking, the first category contains such foulants as calcium carbonate, and
muds and silts that become cemented to the tubing; the second category includes
slimes, algaes, barnacles, etc. Frequently they appear in a combined effect
with the slime, for example, contributing to the deposition of fine, suspended
solids on the exposed surfaces of the condenser and circulating water system.
The buildup of corrosion products on the tube surfaces also acts to reduce heat
transfer efficiency.

The traditional methods for combating fouling and maintaining condenser effi­
ciency are to periodically (or continuously) mechanically remove the deposits,
introduce biocides into the circulating water, and maintain nonscaling water
conditions where such control is necessary and possible.

Of the 12 titanium-tubed units in the field survey, 8 employed the use of

periodic cleaning with brushes, plugs, scrapers, sponge balls, or rods;
4 (including 3 of the 8 above) sprayed the tubing with water jets to remove
accumulations; and 2 cleaned the condensers by reversing flow. All but two of
the units used chlorine for the control of organic growth.

Hydriding/Galvanic Corrosion. A fourth factor relating to the use and perfor­

mance of titanium is the electrical potential differences developed between
titanium and other metals that can result (as explained below) in either: 1) the
adsorption of atomic hydrogen onto the titanium, which may lead to hydrogen
embrittlement; or 2) galvanic wastages of the dissimilar metal materials used

5-52
for tubesheets or water boxes. C. F. Hanson describes the embrittlement
phenomenon (6-p.125):

Atomic hydrogen is formed on eormereially pure titanium surfaces

exposed in seawater when their potential falls below about - 700mV
with respeet to a saturated calomel electrode. Hydrogen evolution
cannot occur when titanium is coupled with any or a combination of
the common copper-based tube or tube plate materials,... It may
occur in systems with sacrifical zinc anodes, or when impressed
current cathodic protection is used. In multi-metal couples
including copper-base materials as well as mild steel, the hydrogen
evolution potential is not reached unless associated with either
impressed current or sacrifical zinc anode cathodic protection.

At the metal temperatures usually encountered in seawater cooled

turbine condensers, surface adsorbed hydrogen atoms normally result
only in discrete surface layer hydriding, which does not unduly
affect the use of the material. It has previously been considered
that diffusion of surface hydride layers into bulk titanium occurs
only above about 80°C (6) (7).* (Diffusion of molecular hydrogen
through the protective titanium oxide film is significant only at
temperatures in excess of Z00-350°C.)

The Japanese have reported adsorption of hydrogen and hydriding of the titanium
below the surface of the tubes after 1 to 2 years of operation in surface con­
densers (135). The titanium was used in the air removal sections of otherwise
aluminum-brass-tubed condensers using uncoated naval brass tubesheets. Ferrous
ion was injected and impressed current cathodic protection was provided. The
reported hydriding occurred on the inside surface in the first 2.4 inches
(60 mm) of length from each end of the tubes to a maximum depth of about
0.025 inch (^0.1 mm). The amount increased toward the ends, and the outlet end
experienced a higher hydrogen adsorption than the inlet end. No failures of
the titanium were reported.

Although details are lacking, the field survey of one titanium-tubed unit
revealed a case of subsurface hydriding on the outside of a tube at (or very
near) a support plate location. The tube had been removed for inspection follow­
ing failure by vibration-induced collision damage and laboratory analysis
revealed hydriding. It is not known whether the hydrogen contamination occurred
before or after installation in the condenser.

*(6) Cotton, J. B., The Science, Technology and Application of Titanium,

Pergamon Press, 1970, 155.
(7) Chariot, L.A., et al. OSW Research and Development, Report No. 624, 1970.

5-53
TIMET (127) has investigated the embrittlement phenomenon in the laboratory by
cathodically charging a variety of titanium samples in a synthetic sea salt
solution at ambient temperatures. In addition, samples of test titanium tubing
from in-service condensers were removed for examination. Their findings are
summarized below:

It appears to be very difficult to charge hydrogen into titanium

or even to form a surface hydride layer at ambient temperature in
a neutral pH solution. Power plant surface condensers operate at
a temperature of about 40°C. At this low temperature^ it seems
highly unlikely that hydrogen embrittlement of the titanium tubes
would occur even if cathodic protection did form a hydride film on
the surface. Jacobs and McMaster (5)* claim that no appreciable
diffusion of hydrogen in titanium occurs at temperatures below 70°C.

Laboratory tests indicate that it is highly unlikely that hydrogen

embrittlement of titanium surface condenser tubes will occur under
normal operating conditions. These findings are supported by the
independent published work of other investigators.

Examination of titanium tubes after years of service in polluted

seawater under cathodic protection support this conclusion. No
evidence of hydrides or deterioration of mechanical properties
could be detected.

Anodizing the surfaces of titanium tubing (electrolytically increasing the oxide

thickness) is used in refinery applications where the material is exposed to
higher temperature hydrogen environments. But where sufficient water is present
to maintain passivity as in steam condenser service, anodizing does not appear
to provide any additional protection (117).

Titanium is cathodic to all alloys typically used for condenser tubesheets and
water boxes. Its use in a condenser increases the normal uncoupled corrosion
rate of any dissimilar metal used in such services. Some likelihood for gal­
vanic corrosion exists for any dissimilar metal couple in condensers. But the
likelihood is higher with titanium because reducing reactions with oxygen and
hydrogen occur more readily on titanium.

*(5) Bibliography reference 117 of this report.

5-54
The Japanese have reported galvanic attack of naval brass tubesheets both with
and without titanium tubing in the air removal sections of condensers. Im­
pressed current cathodic protection was used to control the galvanic attack but
hydriding of the titanium reportedly occurred. The addition of ferrous ion
significantly reduced the galvanic corrosion (135). Other work (133) indicates
that titanium and naval brass can be used together without hydriding or galvanic
attack if the cathodic protection potential is carefully controlled.

The use of coatings and more galvanically compatible materials (i.e., aluminum
bronze tubesheets) have also been suggested to provide protection of materials
coupled with titanium (6-pp.126-127).

The above discussion indicates that impressed current cathodic protection

systems must be well controlled when used with titanium. Table 5-16 tabulates
the data relative to the "galvanic environment" in those condensers tubed with
titanium which were included in the field survey.

Other Corrosion and Erosion Factors. The following is a general summary of the
literature relevant to the susceptibility and resistance of titanium to corrosion
and erosion:

• "The protective oxide film on titanium is maintained in sea­

water at least up to 130°C metal temperature irrespective of
oxygen, sulphide, or ammonia content and under crevice condi­
tions arising from scaling, biofouling, or local tube blockage"
(6-p.124).
• "...commercially pure titanium does not suffer crevice attack
in seawater or in brine at temperatures up to 130°C... An in­
crease in chloride from 3% to 6% does not alter that situation.
Neither does normal fluctuation in the pH value of seawater,
provided that the pH level remains above a figure of about 3"
(6-pp. 121-122) . Similarly, pitting corrosion does not affect
the material below about 250°C in seawater conditions (117;212;
347) .
a "It is immune to stress corrosion, corrosion fatigue and
fretting corrosion in polluted cooling waters and in high
purity chloride contaminated or ammoniated steam" (6-p. 124).
• Due to titanium's extremely low corrosion rate in a variety of
condenser cooling waters and the nontoxic nature of the metal
and its passivating oxide (TiO ), there exist no potential
adverse environmental effects due to its use in cooling water
systems.*

*"Titanium Tubing, Bulletin SC-3," p.13.

5-55
 Table 5-16

FIELD SURVEY DATA

TITANIUM AND ITS GALVANIC ENVIRONMENT

REPORTED REPORTED
TUBESHEET WATERBOX CATHODIC GALVANIC TITANIUM
CONDENSER MATERIAL/COATING MATERIAL/COATING PROTECTION ATTACK* ** HYDRIDING*

A Muntz/none 2% Ni-Ci/none Impressed current No No

B Muntz/none 2% Ni-Ci/none Impressed current No No

C Muntz/none Ci/none Impressed current No No

D Muntz/none steel/gunnite Impressed current No No

E Muntz/none steel/gunnite Impressed current No No

F AlBronze/none steel/carbomastic Sacrificial-zinc No No

G 90-10 Cu-Ni/epoxy steel/epoxy Sacrificial-zinc No Yes - 1 tube from

steamside

H Muntz/none inlet: steel/ Sacrificial-mg Slight on No

fiberglass tubesheet
outlet: Ci/none

I AlBronze/none AlBronze/none Sacrificial-zinc No No

(discontinued
shortly after
TI installation)

J Muntz/none inlet: none None No No

2% outlet: coal
Ni-Ci tar epoxy,
bitumastic

K Muntz/none steel/coal tar None No No

epoxy

L Muntz/none Ci/none None-low TDS No No

water

^Subjective data; e.g., most stations have not specifically checked for hydriding; no
inspection of tubesheets, water boxes, or tubing were made during the survey.

**Severe galvanic attack occurred on some tubesheets prior to the installation of the titanium
tubing. The responses here only apply since the installation of titanium.

5-56
 • Similarly, the steam-side corrosion resistance of titanium
results in minimal metal ion carryover to downstream steam
generating equipment. For BWR applications, the daughter pro­
ducts due to any titanium carryover to the reactor have very
low half lives.*
• Titanium is not adversely affected by conventional acid clean­
ing cycles where it may be installed in mixed material plant
designs with iron or copper-base alloys requiring acid clean-
j_ng. ** ***

• Titanium is resistant to erosion in clean seawater at veloci­

ties up to about 65 fps (20 m/sec); with high sand content
water, the erosion threshold is about 20 fps (6 m/sec)
(6-p.122). Localized tube blockages do not cause erosion
attack at normal design velocities up to 11.5 fps (3.5 m/sec)
(6-p.124).
• Sato (235-p.265) has indicated that some difficulties may exist
with hydrogen attack from the steam side of BWR units and a
probability for crevice corrosion under heavy water-side
deposits. The second point is also brought out by Jacobs and
McMaster (117) that if a solid chloride salt deposit forms on
the titanium surface, the concentrating and insulating effect
of the deposit may increase the surface temperature ("hot wall
effect") and salt concentration sufficiently to cause pitting.

SUMMARY OF PERFORMANCE

The data previously presented in this section under the individual material
headings will now be brought together for comparative analysis. It has been
developed earlier in this report, and also by other investigators, that the
service life of condenser tubes is markedly affected by the composition of the
cooling water. Accordingly, separate comparisons will be made for each of the
three broad cooling water categories used in this study.

Fresh Water

Condensers cooled once-through by fresh water experience significantly less tube

deterioration than do condensers cooled by brackish water or seawater. This is
illustrated by the overall field survey data. The 108 tubesets of all materials
in main-body service in fresh water had a median total failure rate of 0.021.
The 129 tubesets of all materials in main-body service in salt water had a
median failure rate of 0.30 — a fifteenfold increase over that for fresh water.

*Ibid, p.14.
**Ibid, p.16.
***"Titanium Tubes for Refinery Heat Exhangers, Bulletin RT-1," Titanium
Metals Corporation of America, November 1975, p.21.

5-57
A comparative summary of tube deterioration measurements for the principal con­
denser tube materials used in fresh water is given in Table 5-17. The data base
for each category of material and service is described by the number of stations,
generating units, and tubesets for which failure rates could be calculated. In
main-body service, the broadest data base is for admiralty; SS-304 is interme­
diate; and 90-10 Cu-Ni has the narrowest base with all data from a single sta­
tion. In air removal service, an adequate data base was obtained only for
admiralty and SS-304.

Main-Body Service. The tube deterioration parameters listed in Table 5-17

indicate that in the main condensing body of freshwater cooled condensers,
admiralty, SS-304, and 90-10 Cu-Ni all give acceptable service life. There is
no outstanding difference among them; the variations in the tabulated values
are small enough to be less than the probable error of the stated values. For
all three materials the probability (calculated from the respective data bases)
is greater than 90 percent that the total failure rate due to all causes will
not exceed 0.333. This corresponds to the failure rate that would result if
10 percent of the tubes failed after 40 years of service or if 5 percent failed
after 20 years. The calculations for SS-304 were adjusted as noted in Table
5-17. Other data bases were not adjusted similarly because the number of
guard plugs present was insufficient to dominate the calculation and cause a
possibly misleading result.

It is noteworthy that of the 108 tubesets in main-body service, only two

(1.85 percent) have been replaced. The two replaced tubesets were parallel
bundles in the same condenser.

Although the anticipated service life of the principal condenser tube alloys in
freshwater service may be approximately equal, the selection of tube material for
any specific application must carefully consider numerous other factors. In
addition to such ordinary concerns as heat transfer capability, tubing availa­
bility, economics, etc, there has been a recent concern about the release of
copper and other metals from the condenser tubes to the condensate, and the
impact of the released metals on the turbine cycle equipment downstream of the
condenser. Evaluation of the release of metals from the condenser tubes to the
condensate was not a part of this study.

5-58
 Table 5-17

COMPARATIVE PERFORMANCE OF CONDENSER TUBE MATERIALS

IN OPEN FRESHWATER COOLING SYSTEMS

PROBABILITY, %
NUMBER OF TUBESETS REPLACED MEDIAN TOTAL 40YR - 10% 30YR - 5%
SERVICE MATERIAL STATIONS UNITS TUBESETS NUMBER PERCENT FAILURE RATE FR <0.333 FR <0.222

Main Body Admiralty 14 25 66 2 3.0 0.031 91 88

Main Body 90-10 Cu-Ni 1 2 16 0 0 0.045 100 81
Main Body SS-304 5 6 26 0 0 0.000* 96* 96*
6S-

Air Removal Admiralty 5 9 34 20 58.8 3.398 27 15

Air Removal 90-10 Cu-Ni 1 2 4 0 0 ** ** **

Air Removal SS-304 9 14 42 0 0 0 93 93

Air Removal SS-316 1 1 4 0 0 0 ** **

Air Removal Titanium 1 1 4 0 0 0 ** **

*These values have been adjusted by the elimination of numerous guard plugs installed in four tubesets at one
station. The raw data values are: Median Failure Rate = 0.003, 40yr - 10% probability = 80.8%, 30yr - 5%
probability = 80.8%.

**There are not sufficient values in the data base to allow a meaningful calculation to be made.
Figure 5-12 is a plot of the total failure rate versus the operating time for
the tubesets in main-body service in freshwater cooled condensers. The material
is designated by the indicated symbol. Twenty admiralty tubesets and 16 SS-304
tubesets had failure rates less than 0.01 and are not shown on the plot. The
only value plotted for each tubeset is the cumulative total failure rate (all
causes) for the total operating hours of the tubeset at or near the date of the
field survey. Examination of Figure 5-12 shows no correlation between the
attained failure rate and the age of the individual tubeset for any of the
three tube materials.

To place this observation in proper perspective it should be recalled that the

total failure rate is calculated from the number of plugged tubes in the con­
denser without regard to why the tubes were plugged. It includes guard plugs
deliberately placed in tubes that had not failed at the time. While it is a
practical measure of condenser tube serviceability because condenser performance
and maintenance costs relate to the number of tubes plugged and not to the
reason for their plugging, it does not indicate, in any way, the failure
mechanism. The failure rate due to some mechanisms such as electrochemical
corrosion would be expected to increase exponentially with time. The lack of a
correlation between the attained total failure rate and the tubeset service life
suggests that no one failure mechanism is dominant overall. Apparently, the
net effect of all failure mechanisms working together is independent of time
when a large body of actual historical data is considered as is done in Figure
5-12.

It was also found that no overall correlation existed when the total failure
rates (FR) for the individual tubesets were plotted against the selected time
increments (Field Survey Matrix columns 117 through 124) up to the attained
service life (only these last points are shown in Figure 5-12) . Also, no
overall correlation was found when the available data for failure rate on the
water-side (FRW) was plotted against time. (These data are given in the Field
Survey Matrix colums 93 through 100.)

The line labeled FRft = 100// H is a criterion for judging if a given tubeset
is a candidate for replacement. It is discussed on page 5-66.

Air Removal Service. As shown in Table 5-17 for the tubes in air removal ser­
vice, an adequate data base was obtained for admiralty and SS-304 only.

5-60
1000
X A
800
X^ FRESHWATER - MAIN BODY

600 X
©Admiralty 66Tubesets

□ SS304 26Tubesets
400
A 90-10 Cu-Ni 16Tubesets
Total 108 Tubesets
300
X In Symbol indicates Replaced Tubeset
Superscript indicates the number
© o of identical values
200 w
O2 ©
0X \ FR 10C \^V
0
X Gh
© © 0 o
4 02 0
C)
©
100
w 0 © _£
80
© 1®
60 fry iO

□ Ld
@22
40 0) O G) G)
(ill) ©

A _^2 A .2.
20

A3 A
2
1 oo
3e X
10
0.02 0.05 0.1 0.2
TOTAL FAILURE RATE
0.5 1.0
\ 2.0 5.0

Figure 5-12. Failure Rate vs Operating Time for Main-Body Tubesets

in Fresh Water
The tabulated performance criteria indicate that admiralty gives unsatisfactory
service in the air removal section of freshwater cooled condensers. For the
34 tubesets surveyed, the median failure rate was 3.398. Nearly 59 percent of
the tubesets have been replaced. The dominant failure mode apparently was
steam-side failures induced by ammonia.

In the air removal section, SS-304 gives very acceptable performance. All of
the 42 tubesets surveyed are still in service with a median failure rate of
zero. Based on these data there is a 93 percent probability that SS-304 in air
removal sections will serve 40 years with less than 10 percent of the tubes
plugged.

Salt Water

In Table 5-18 the comparative summary of the measures of tube deterioration are
given for the principal materials used in condensers cooled once-through by salt
or brackish water.

Main-Body Service. In this service a reasonably adequate data base was obtained
for aluminum brass, 90-10 Cu-Ni, and titanium. Lesser data bases were obtained
for 70-30 Cu-Ni and aluminum bronze. Viewing the main-body data as a whole it;
appears that the aluminum brass, aluminum bronze, and 90-10 Cu-Ni will not last
the plant's lifetime.

Based on the survey data, all three alloys have a 50 percent or less probability
of attaining a failure rate less than 0.33 (40 years with less than 10 percent of
the tubes plugged). There is a 40 percent or less probability that the failure
rate will be less than 0.222 which corresponds to 30 years service with less
than 5 percent of the tubes plugged. Of the surveyed tubesets, a high propor­
tion had been replaced: 45 percent of the 90-10 Cu-Ni, 60 percent of the
aluminum brass, and 67 percent of the aluminum bronze.

The data base obtained for 70-30 Cu-Ni consisted of only 6 tubesets in main body
service, which is inadequate to establish a meaningful trend. The very limited
data on 70-30 Cu-Ni indicate a somewhat better performance than that for
aluminum brass, aluminum bronze, and 90-10 Cu-Ni.

5-62
 Table 5-18

COMPARATIVE PERFORMANCE OF CONDENSER TUBE MATERIALS

IN OPEN SALT-WATER COOLING SYSTEMS

PROBABILITY, %
NUMBER OF TUBESETS REPLACED MEDIAN TOTAL 40YR - 10% 30YR - 5%
SERVICE MATERIAL STATIONS UNITS TUBESETS NUMBER PERCENT FAILURE RATE FR <0.333 FR <0.222

Main Body Alum. Bronze 2 3 12 8 66.7 1.880 33 33

Main Body Alum. Brass 6 11 35 21 60.0 0. 333 50 40
-63

Main Body 90-10 Cu-Ni 7 16 55 25 45.5 0.613 40 29

Main Body 70-30 Cu-Ni 2 2 6 2 33.3 0.102 67 67
Main Body Titanium 8 11 21 0 0 0.003 91 91

Air Removal 70-30 Cu-Ni 5 11 45 8 17.8 0.060 69 62

Air Removal Titanium 1 1 2 0 0 0.000 * *

*There are not sufficient values in the data base to allow a meaningful calculation to be made.
The data base for titanium tubes in main-body service in condensers cooled once-
through by brackish water or seawater comprised 21 tubesets. As previously
stated, nearly all United States power stations with a significant number of
titanium condenser tubes were included in the survey hence the data base is
nearly the sum of experience in the United States. As shown in Table 5-18,
titanium gives excellent service in seawater service. The calculated probabil-
lity is 91 percent that the failure rate will be less than 0.333 (40 years with
less than 10 percent plugs). This probability would be higher (approaching
99 percent) if it were not for a number of vibration failures (and affiliated
guard plugs) at one station where the titanium tubes were installed as replace­
ments in an existing condenser.

The median failure rate for the titanium data base was 0.003. As pointed out in
the discussion of titanium in this section of the report, the oldest titanium
tubesets are only 5 years old. Prudent judgment should be exercised in the
application of the implied extrapolations given by the probability calculation
interpretations.

Total failure rates versus the operating time to the survey date are plotted in
Figure 5-13 for all surveyed tubesets in main-body service in condensers cooled
once-through by salt or brackish water. The material is designated by the
indicated symbol. An X through the material symbol indicates that the tubeset
has been replaced.

As was the case for fresh water (Figure 5-12), there is no discernible correla­
tion between the total failure rate attained as of the survey date and the age
of the tubeset. The comments on this observation given in the discussion of
Figure 5-12 are applicable here also.

Examination of Figure 5-13 allows insight as to when the industry deems it

desirable to replace the tubes in a failing condenser. Nearly all of the
replaced tubesets had an attained failure rate greater than 0.222 at the time
of replacement. Of replaced tubesets that had a failure rate less than 0.222,
several were in condensers where other tubesets had a failure rate greater than
0.222 and the entire condenser was retubed. Most of the replaced tubesets had
a failure rate greater than 0.333 at the time of replacement. Note that nearly

5-64
 TUBES PLUGGED
5% _________10%_________ 20%
1000

800 SEAWATER - MAIN BODY

V 70-30 Cu-Ni 6 Tubesets
600
□ Aluminum Brass 35 Tubesets
A Aluminum Bronze 12 Tubesets
400 0 90-10 Cu-Ni 55 Tubesets
A Titanium 21 Tub'esets
300 Total 129 Tubesets
X In Symbol indicates Replaced Tubeset
Superscript indicates the number
, of identical values
200

\FRA =—
RATING TIME, Khi

P0 \ ^
F \
r \
100

80
□ □ c ) 03 \ w
...... £ □ LJ La!
El \
a! 60 * dXV El V KII5 3___
O A A A
\ 0 Si >30 ft a___ X
El A AA Su-,
40

) ©
0
El « G
® \ ®2 \V
>
[] □ 02 \ \
o 0
\
A X E □ g) s
<

20
y V " V
-----------------

© 0 0
© o 00

El
E 3 El2
N3

(7) )

L
10
0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
FAILURE RATE

G1 001892-30

Figure 5-13. Failure Rate vs Operating Time for Main-Body Tubesets in Seawater
all of the replaced tubesets had more than 1 percent of the tubes plugged, and
most had more than 2 percent plugged. About half of the replaced tubesets had
5 percent or more plugged tubes but only a few had more than 10 percent of the
tubes plugged. However, the attained failure rate at the time for replacement
is not an adequate criterion for tracking the performance of an in-service tube-
set. For example: Early in the life of a tubeset a number of tube failures
could be experienced due to vibration or due to a missile. The calculated
failure rate would be high due to the short operating time even though the
percentage of the tubes plugged was less than 2 percent. After correction of
the mechanical problem the calculated failure rate would decrease as the opera­
ting hours increase with few, if any, additional failed tubes. Hence a more
valid criterion for assessment of the performance of an in-service tubeset would
be a variable failure rate. The "allowable" failure rate would decrease as the
service life of the tubeset increased. A line that accomplishes this is drawn
in Figure 5-13. The equation of the plotted line is:

100
/l?
Where FR = allowable failure rate
A
H = total operating hours for the tubeset

The equation of the line can also be expressed as:

%P
A 100

Where %PA = allowable % tubes plugged

H = total operating hours for the tubeset

This line was fitted with the intent that failure rates lying to the right of
it would be candidates for replacement while failure rates to the left of the
line would indicate a high probability that the replaced tubeset would serve
the plant's lifetime. In Figure 5-13, 72 percent of the tubesets lie to the
right of the line, and 28 percent lie to the left. Recall, however, that about
half of the replaced tubesets to the left were in condensers with other tube-
sets that were to the right of the line.

5-66
A more conservative criterion can be obtained by multiplying the calculated
allowable failure rate by an appropriate factor. For example: If the calculated
FR^ is multiplied by 0.5 then 85 percent of the failed tubesets in Figure 5-13
would be to the right of the more conservative line. All of the tubesets
remaining on the left are in condensers with other tubesets to the right. The

The use of either equation provides a simple and rapid method for tracking the
performance of existing condenser tubesets based on the cumulative data obtained
during this study.

Plant management can readily determine when the tubeset is a candidate for re­
placement by comparing the attained failure rate with the allowable failure rate.

Air Removal Service. The only material in air removal service in salt-water-
cooled condensers for which an adequate base was obtained was 70-30 Cu-Ni. The
field survey included 42 tubesets of 70-30 Cu-Ni. The median failure rate was
0.060. Nearly 18 percent of the tubesets had been replaced. The calculated
probability of attaining a failure rate less than 0.333 (40 years, 10 percent)
was 69 percent.

Closed Cooling-Water Systems

The comparative summary of the measures of tube deterioration for condenser tubes
in closed cooling-water systems is given in Table 5-19. All of the closed
systems included in the survey used fresh water for cooling tower makeup.

The data base for all materials is small. Accordingly, the indicated perfor­
mance parameters have limited reliability. There is an indication that admiralty
in main body service gives much poorer performance in a closed cooling system
than it does in a freshwater once-through cooling system. On the basis of the
limited sample, SS-304 in main body service appears to perform as well in a
closed cooling-water system as it does in a freshwater once-through system.

In the air removal section of condensers with closed cooling-water systems,

70-30 Cu-Ni appears to give very acceptable service. Here again, the data base
is quite limited.

5-67
 Table 5-19

COMPARATIVE PERFORMANCE OF CONDENSER TUBE MATERIALS

IN CLOSED COOLING-WATER SYSTEMS

PROBABILITY, %
NUMBER OF TUBESETS REPLACED MEDIAN TOTAL 40YR - 10% 30YR - 5%
SERVICE MATERIAL STATIONS UNITS TUBESETS NUMBER PERCENT FAILURE RATE FR <0.333 FR <0.222

Main Body Admiralty 3 6 16 2 12.5 0.135 56 56

Main Body SS-304 2 3 9 0 0 0.107 89 89
-68

o
o
Air Removal 70-30 Cu-Ni 2 2 10 0 0 90 90
Air Removal 90-10 Cu-Ni 1 2 4 0 0 3.376 * •k

Air Removal SS-304 2 5 8 - - * * *

*There are not sufficient values in the data base to allow a meaningful calculation,to be made.
 Section 6

CONCLUSIONS

The conclusions listed in this section are derived from an objective evaluation
of the findings of this study. They are based primarily on the data obtained
by the field survey but they also reflect the findings of the mail survey and
the literature search. Other information sources can be expected to produce
specific exceptions to these conclusions, but because of the methodology used
in the selection of the field survey samples (Section 3) and the size of the
sample (76 generating units at 30 stations), these conclusions should be repre­
sentative of condenser related experience at large, modern power stations in the
U.S.

PERFORMANCE OF CONDENSER TUBE MATERIALS

Fresh Water

1. In the main condensing section of condensers cooled by fresh water in a

once-through mode, all of the commonly used condenser tube materials give
acceptable service. This has been shown in Table 5-17 which summarizes the
performance data collected by the field survey for admiralty, 90-10 Cu-Ni
(Alloy 706), and Type 304 stainless steel. For these materials there is
a better than 90 percent probability that the total failure rate will not
exceed 0.333 which corresponds to 10 percent of the tubes plugged after
40 years of service or 5 percent plugged after 20 years.

2. Complete retubing of freshwater cooled condensers is seldom required. Only

2 of 108 (1.85 percent) tubesets in main body, freshwater service have been
retubed (Table 5-17).

3. In the condensing section of freshwater cooled condensers, the failure rate

from all causes (plugged tubes for whatever reason) does not correlate with
the age of the individual tubeset (see Figure 5-12).

6-1
4. In the air removal section of condensers cooled by the once-through flow
of fresh water, some severely high failure rates have been encountered when
admiralty tubes have been used. In this service, the probability that
admiralty will have a total failure rate less than 0.333 (10 percent plugs
in 40 years) is only 26.5 percent (Table 5-17). The apparent reason for
this poor performance is the presence of ammonia.

5. In the air removal section of freshwater cooled condensers, SS-304 has a

96.2 percent probability of serving 40 years with less than 10 percent
failures (Table 5-17) .

6. In the air removal section also, no correlation can be made between the
failure rate from all causes and the age of the tubeset.

7. Insufficient data were gathered to allow an evaluation of 90-10 Cu-Ni

tubing in air removal section service.

Cooling Towers and Ponds

8. The very limited data bank for closed cooling-water cycle condensers does
not allow formulation of valid conclusions. The available data indicate
that in the condensing section admiralty may have a lessened service life
as compared to open cycle freshwater applications. Type 304 stainless
steel performs almost as well in closed cooling water cycles as in
fresh water (see Table 5-19).

Salt Water

9. Until recently, the traditional condenser tube alloys used in the conden­
sing section of saltwater cooled condensers were aluminum brass and
aluminum bronze. Neither alloy can be expected to last the plant lifetime
without retubing. The probability of 40 years service with less than 10
percent tube failures is 33.3 percent for aluminum brass and 50 percent
for aluminum bronze (see Table 5-18).

10. Alloy 706 (90-10 Cu-Ni) has been used in the condensing section of salt­
water cooled condensers in recent years. Its anticipated service life is
about the same as for aluminum brass/bronze alloys. The probability that
90-10 Cu-Ni will serve 40 years with less than 10 percent tube failures is
40 percent (see Table 5-18).

6-2
11. Titanium, the newest condenser tube material, is the only material included
in this survey that can be expected to last the plant lifetime in the con­
densing section of saltwater cooled condensers. Titanium was found to
have a 90 percent probability of lasting 40 years with less than 10 percent
tube failures. This probability would be nearly 100 percent if it were not
for guard plugs placed in 2 tubesets that were experiencing vibration
problems.

12. All of the actual failures of titanium tubes were attributable to mechani­
cal causes, predominantly vibration. Not one instance of corrosion or
erosion failure was found during the field survey.

13. In air removal service in saltwater cooled condensers, a significant data

base was generated for only 70-30 Cu-Ni. In such service, 70-30 Cu-Ni has
a 69 percent probability of serving 40 years with less than 10 percent tube
failures (see Table 5-18).

14. In saltwater service (as in freshwater service), there is no correlation

between the total failure rate from all causes and the age of the tubeset
(see Figure 5-13) .

CONDENSER RETUBING CRITERIA

1. No single criterion can be expected to define the time when tubes in a

failing condenser should be replaced.

2. The primary considerations in the decision to retube relate to the cost of

repairing the leaks as they occur, and to the attendant loss in the relia­
bility and full availability of the generating unit. The possibility of
damage to other components in the turbine cycle must also be evaluated.

3. From the data accumulated during this study a numerical criterion has been
derived to indicate when the condenser is a candidate for retubing. The
criterion reflects the historical judgment and practice of the industry.
When the actual failure rate (FR) exceeds the allowable failure rate (FRA)
it is likely that the condenser should be retubed. The final decision to
retube must also consider the factors described in 2. above. The equation
empirically derived for allowable failure rate is:

6-3
 where, H = total operating hours for the existing condenser tubes
FR = allowable failure rate
A
This can also be expressed as:

o. _ Zji
*PA 100

where, %PA = allowable percent of condenser tubes plugged

The actual failure rate FR is calculated as follows:

T

„ P/N x 106
FR = —-------
T H

where, P = number of tubes plugged

N = number of tubes in tubeset

IMPACT OF CONDENSER LEAKAGE

1. At the generating units surveyed there was no demonstrated case of severe

physical damage to any of the other components in the turbine cycle as a
result of condenser inleakage.

2. At at least one station, the possibility that condenser inleakage was con­
tributory to serious damage to the steam generator tubes is still under
consideration.

3. The major impact of condenser tube leakage is the value of the power
generation lost while the unit is operated at reduced load or is out of
service.

4. No generally applicable cost figures can be assigned. The costs vary

widely at the different utility systems largely because of variations in
the value of the generation lost during repair operations.

A condenser with failing tubes can have an unacceptable impact on the

availability and reliability of the generating unit.

6-4
6. The loss of condenser surface area due to plugged tubes does not have a
major impact on the turbine capability. Calculations for a typical
fossil-fueled power plant with a 720 MWe turbine show the following:

Tubes Plugged Reduction in Turbine Output

1% 50 kW or 0.007%
2% 100 kW or 0.014%
5% 250 kW or 0.035%
10% 540 kW or 0.075%

DETECTION AND LOCATION OF CONDENSER LEAKS

1. All plants surveyed can detect any condenser leak large enough to be of
consequence to plant operation by analysis of the condensate and steam
generator blowdown.

2. Most plants have automatic analyzers to monitor turbine cycle chemistry

which will respond adequately to condenser leaks of consequence.

3. The actual leak size that can be detected depends upon the quality of the
circulating water. Sea water cooled plants can detect leaks of a few
tenths of a gpm. Plants with low dissolved solids in the circulating
water may not be able to detect a leak smaller than 10 gpm.

4. Detection methods (especially for plants with high TDS circulating water)
are usually more sensitive than methods for locating the leak within the
condenser. However, all plants can locate any leak large enough to be of
consequence to continued plant operation.

5. There is general agreement at the surveyed stations that the most sensitive
method for leak location is to hydrotest the condenser on the steam side.6

6. For on-line location of condenser leaks two of the methods reported to be

more sensitive are the spreading of either plastic sheets or foam over the
tubesheet of the open water box. There is no general agreement as to which
method is the most sensitive.

6-5
 Section 7

RECOMMENDATIONS

The recommendations drawn from the findings and conclusions of this study are
presented in three groups based on the time period recommended to accomplish
the work: immediate short-term work; middle-term work; and long-term work.

SHORT-TERM WORK

1. Gather and evaluate additional data to enable a better assessment

to be made of the performance of 90-10 Cu-Ni in the main body of
condensers cooled by once-through fresh water and, most importantly,
in condensers cooled by closed cooling water systems.

3. Make quantitative investigations of the copper pickup in the con­

densate from condensers tubed with the copper alloys. These data
are urgently needed because of the current concern about the role
of copper in the failure of Inconel tubes in PWR steam generators.
The data are also important relevant to copper deposition in fossil-
fueled boilers.

3. Develop a program to promote rigorous design standards for impinge­

ment baffles in condensers. Far too many condenser tube leaks,
especially the larger ones, are caused by failed baffles that either
sever the tubes directly or allow their rapid failure due to steam
impingement.

4. Investigate methods to reduce the number of drain, dump, etc, pene­

trations into condensers and thereby reduce the need for impingement
baffles. Present-day condensers have 100 or more penetrations,
each of which is also a potential source of air inleakage.

5. Develop and promulgate a standard tube descriptor or numbering

system for condenser tubes.

7-1
6. Further investigate the frequency and size of leaks at the tube-to-tube-
sheet joint. The plants surveyed in this study reported few such leaks.
Joint leaks are usually small in size and precise location techniques
are cumbersome and time consuming. The need for improved joining methods
can be better assessed if more specific data are available.

7. Further investigate the role of condenser design parameters as a cause of

condenser leaks. This survey did not produce sufficient data on the
specific cause of leaks to permit evaluation of the effect, if any, of
the shell, tubesheet, and water box configurations, the tube support plate
spacing, the strength of tube joints, and the design of expansion joints.

8. Additional evaluation is desirable on the cost effectiveness of such tech­

niques for possible reduction of condenser inleakage as double tubesheets
and welded tube joints.

9. Begin Phase I of the long-term investigation.

MIDDLE-TERM WORK

10. Develop and promulgate a manual on proper record-keeping relative

to condenser performance that would include standardized log sheets
and other appropriate forms.

11. Initiate a study on the reduction of penetrations to the condenser.

Perhaps a separate drain/flash tank should be provided for all dump,
drain, etc, lines which would have a single large connection to the
condenser. The condenser shell should not be used as a piping mani­
fold to the extent that it presently is.

12. Complete Phase II of the proposed long-term study.

LONG-TERM STUDY

13. The present study clearly indicates the need for a continuing long­
term (10 years, or more) study of the general subject area included
in this report. As indicated in the report, the data base obtained
in the course of the brief field survey program was inadequate to
enable definitive evaluation of some condenser tube alloys in cer­
tain environments. Also, it was found to be next to impossible to

7-2
 reconstruct the data and events surrounding leakage after the
occurrence even when the requisite data were in the plant archives.
The data obtained on the frequency, size, and duration of condenser
leaks were quite limited.

The obvious solution is a self-sustaining program which would cap­

ture and report the necessary data at the time of the incident,
recording such data as leak size, subsequent impact, cost of
repair, and causative mechanism. Clearly, such a program should
place major emphasis on seawater-cooled stations since this is
"where the action is." The program should be designed to enable
plant personnel to report data on a semiannual basis. The task
would be simplified by the use of standard forms to be completed
routinely and problem alert forms for special incidents. The in­
coming data would be entered in a computerized data bank on receipt.
The accumulated data would be analyzed and reported at annual intervals.
The evaluation, design, and implementation of the program would be in
three phases. Each phase is described below by a listing of the principal
tasks that would constitute the phase.

Phase I. Feasibility and Cost Effectiveness

a. Establish the data to be collected.

b. Devise preliminary forms for data collection.
c. Determine mode and extent of data manipulation and reports to be
prepared.
d. Make cost estimates of the startup cost and the annual operating
cost for the program.
e. Determine if program is feasible and cost-beneficial. If not,
abandon.

Phase II. Program Design

a. Prepare detailed data forms, etc.

b. Devise necessary computer programs.
c. Prepare conservative estimate of project costs.
d. Prepare preliminary list of survey stations.
e. Prepare detailed instructions for station participation.

7-3
Phase III. Implementation

a. Select and organize responsible agency.

b. Prepare final station survey list (^100 stations).
c. Complete program design details.
d. Collect available initial data and load into computer.
e. Distribute field forms to participating stations.
f. Enter data semiannually.
g. Prepare annual reports.
h. Review progress each year and adjust program accordingly.

7-4