Te 981 PRN PDF

IAEA-TECDOC-981
Assessment and
management of ageing of major
nuclear power plant components
important to safety:
Steam generators
INTERNATIONAL ATOMIC ENERGY AGENCY

The IAEA does not normally maintain stocks of reports in this series.
However, microfiche copies of these reports can be obtained from
IN IS Clearinghouse
International Atomic Energy Agency
Wagramerstrasse 5
P.O. Box 100
A-1400 Vienna, Austria
Orders should be accompanied by prepayment of Austrian Schillings 100,

in the form of a cheque or in the form of IAEA microfiche service coupons
which may be ordered separately from the IN IS Clearinghouse.
The originating Section of this publication in the IAEA was:
Engineering Safety Section
International Atomic Energy Agency
Wagramerstrasse 5
P.O. Box 100
A-1400 Vienna, Austria
ASSESSMENT AND MANAGEMENT OF AGEING OF MAJOR

NUCLEAR POWER PLANT COMPONENTS IMPORTANT TO SAFETY:
STEAM GENERATORS
IAEA, VIENNA, 1997
IAEA-TECDOC-981
ISSN 1011-4289
©IAEA, 1997
Printed by the IAEA in Austria

November 1997
FOREWORD
At present, there are over four hundred operational nuclear power plants (NPPs) in IAEA
Member States. Operating experience has shown that ineffective control of the ageing
degradation of the major NPP components (e.g. caused by unanticipated phenomena and by
operating, maintenance, design or manufacturing errors) can jeopardize plant safety and also
plant life. Ageing in NPPs must be therefore effectively managed to ensure the availability
of design functions throughout the plant service life. From the safety perspective, this means
controlling within acceptable limits the ageing degradation and wearout of plant components
important to safety so that adequate safety margins remain, i.e. integrity and functional
capability in excess of normal operating requirements.
This TECDOC is one in a series of reports on the assessment and management of ageing
of the major NPP components important to safety. The reports are based on experience and
practices of NPP operators, regulators, designers, manufacturers, and technical support
organizations and a widely accepted Methodology for the Management of Ageing of NPP
Components Important to Safety which was issued by the IAEA in 1992. They have been
compiled using contributions from technical experts in typically 10 to 12 countries for each
report, a feedback from a September 1994 Technical Committee Meeting attended by 53
technical experts from 21 Member States (whoreviewed first drafts in specialized working
groups), and review comments from invited specialists.
The current practices for the assessment of safety margins (fitness-for-service) and the
inspection, monitoring and mitigation of ageing degradation of selected components of Canada
deuterium-uranium (CANDU) reactor, boiling water reactor (BWR), pressurized water reactor
(PWR), and water moderated, water cooled energy reactor (WWER) plants are documented
in the reports. These practices are intended to help all involved directly and indirectly in
ensuring the safe operation of NPPs; and also to provide a common technical basis for
dialogue between plant operators and regulators when dealing with age-related licensing
issues. Since the reports are written from a safety perspective, they do not address life or life-
cycle management of the plant components, which involves the integration of ageing
management and economic planning. The target audience of the reports consists of technical
experts from NPPs and from regulatory, plant design, manufacturing and technical support
organizations dealing with specific plant components addressed in the reports.
The component addressed in the present publication is the steam generator of the
CANDU, PWR and WWER nuclear power plants. The contributors to the drafting and review
of this TECDOC are identified at the end of this publication. Their work is greatly
appreciated, hi particular, the contributions of P.E. MacDonald, C. Maruska and V.N. Shah
are acknowledged. The officer who directed the preparation of the report was J. Pachner of
the Division of Nuclear Installation Safety.
EDITORIAI. NOTE
In preparing this publication for press, staff of the IAEA have made up the pages from the
original manuscript(s). The views expressed do not necessarily reflect those of the IAEA, the
governments of the nominating Member States or the nominating organizations.
Throughout the text names of Member States are retained as they were when the text was
compiled.
The use of particular designations of countries or territories does not imply any judgement by
the publisher, the IAEA, as to the legal status of such countries or territories, of their authorities
and institutions or of the delimitation of their boundaries.
The mention of names of specific companies or products (whether or not indicated as
registered) does not imply any intention to infringe proprietary rights, nor should it be construed
as an endorsement or recommendation on the part of the IAEA.
CONTENTS
1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2. Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3. Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4. Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2. STEAM GENERATOR DESCRIPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1. PWR recirculating steam generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.2. CANDU recirculating steam generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3. PWR once-through steam generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4. WWER steam generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3. STEAM GENERATOR DESIGN BASIS, FABRICATION AND MATERIALS . . . . 18
3.1. Codes and specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

3.2. Fabrication and materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.1. Heat exchanger tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.2. Tube installation in the tubesheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2.3. Tube supports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.2.4. Feedwater nozzle and shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4. STEAM GENERATOR DEGRADATION MECHANISMS . . . . . . . . . . . . . . . . . . . . 28
4.1. PWR and CANDU recirculating steam generator tubes . . . . . . . . . . . . . . . . . . . . 28

4.1.1. Primary water stress corrosion cracking (PWSCC) . . . . . . . . . . . . . . . . . 29
4.1.2. Outside diameter stress corrosion cracking (ODSCC) . . . . . . . . . . . . . . . 36
4.1.3. Fretting, wear and thinning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.1.4. Pitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.1.5. Denting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.1.6. High-cycle fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.1.7. Wastage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.2. PWR once-through steam generator tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.2.1. Erosion-corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.2.3. Low-temperature primary-side stress corrosion cracking . . . . . . . . . . . . 45
4.2.4. Outside diameter intergranular stress corrosion cracking (IGSCC)
and intergranular attack (IGA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.3. WWER steam generator tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.4. Tube rupture events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.4.1. Tube ruptures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.4.2. Incipient tube rupture events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.5. PWR and CANDU steam generator shell, feedwater nozzle and tubesheet . . . . . 52
4.5.1. Corrosion-fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.5.2. Transgranular stress corrosion cracking . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.5.4. Erosion-corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.6. WWER collector, shell, and feedwater distribution system . . . . . . . . . . . . . . . . . . 57
4.6.1. Stress corrosion cracking of the WWER-1000 collectors . . . . . . . . . . . . 57
4.6.2. Erosion-corrosion of the feedwater distribution system . . . . . . . . . . . . . . 59
4.6.3. Failure of collector cover bolts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.7. Summary of current world experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5. STEAM GENERATOR AGEING MANAGEMENT:

OPERATIONAL GUIDELINES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.1. Primary coolant system water chemistry control parameters . . . . . . . . . . . . . . . . . 66

5.2. Secondary coolant system water chemistry control parameters . . . . . . . . . . . . . . . 69
5.3. Measures to control secondary-side chemical impurity incursions . . . . . . . . . . . . 77
5.4. Measures to remove secondary-side impurities . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.5. Measures to control steam generator deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6. STEAM GENERATOR INSPECTION AND MONITORING REQUIREMENTS

AND TECHNOLOGIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.1. Tubing inspection requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

6.1.1. Tubing inspection requirements in the USA . . . . . . . . . . . . . . . . . . . . . . 82
6.1.2. Tubing inspection requirements in Canada . . . . . . . . . . . . . . . . . . . . . . . 87
6.1.3. Tubing inspection requirements in the Czech Republic . . . . . . . . . . . . . . 88
6.1.4. Tubing inspection requirements in France . . . . . . . . . . . . . . . . . . . . . . . . 88
6.1.5. Tubing inspection requirements in Germany . . . . . . . . . . . . . . . . . . . . . . 88
6.1.6. Tubing inspection requirements in Japan . . . . . . . . . . . . . . . . . . . . . . . . . 88
6.1.7. Tubing inspection requirements in Russia . . . . . . . . . . . . . . . . . . . . . . . . 89
6.1.8. Tubing inspection requirements in Slovenia . . . . . . . . . . . . . . . . . . . . . . 89
6.1.9. Tubing inspection requirements in Spain . . . . . . . . . . . . . . . . . . . . . . . . . 89
6.1.10. Tubing inspection requirements in Sweden . . . . . . . . . . . . . . . . . . . . . . . 89
6.1.11. Tubing inspection requirements in Switzerland . . . . . . . . . . . . . . . . . . . . 90
6.1.12. EPRI tubing inspection recommendations . . . . . . . . . . . . . . . . . . . . . . . . 90
6.2. Tubing inspection techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6.2.1. Eddy current testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6.2.2. Ultrasonic and other inspection methods . . . . . . . . . . . . . . . . . . . . . . . . . 97
6.2.3. Destructive testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
6.3. Monitoring leakage from tubing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
6.4. Feedwater nozzle inspection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
6.5. Monitoring fatigue damage to feedwater nozzles . . . . . . . . . . . . . . . . . . . . . . . . . 102
7. STEAM GENERATOR ASSESSMENT METHODS AND FITNESS FOR

SERVICE GUIDELINES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
7.1. Tubing repair criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

7.1.1. Generic fitness-for-service criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
7.1.2. Defect type and location specific repair criteria . . . . . . . . . . . . . . . . . . . 106
7.2. Tubing fitness-for-service guidelines in various countries . . . . . . . . . . . . . . . . . Ill
7.2.1. Regulatory practices and fitness-for-service guidelines in
the USA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ill
7.2.2. Alternative US fitness-for-service guidelines for outside
diameter IGSCC/IGA at tube support plates . . . . . . . . . . . . . . . . . . . . . . 113
7.2.3. Alternative US fitness-for-service guidelines for PWSCC
in the roll transition region proposed by EPRI . . . . . . . . . . . . . . . . . . . . . 116
7.2.4. Other alternative fitness-for-service guidelines in the USA . . . . . . . . . . . 118
7.2.5. Regulatory practices and fitness-for-service guidelines in Belgium . . . . 118
7.2.6. Regulatory practices and fitness-for-service guidelines in Canada . . . . . 120
7.2.7. Regulatory practices and fitness-for-service guidelines
in the Czech Republic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
7.2.8. Regulatory practices and fitness-for-service guidelines in France . . . . . . 123
7.2.9. Regulatory practices and fitness-for-service guidelines in Germany . . . . 125
7.2.10. Regulatory practices and fitness-for-service guidelines in Japan . . . . . . . 125
7.2.11. Regulatory practices and fitness-for-service guidelines in Russia . . . . . . 126
7.2.12. Regulatory practices and fitness-for-service guidelines in Slovenia . . . . 126
7.2.13. Regulatory practices and fitness-for-service guidelines in Spain . . . . . . . 127
7.2.14. Regulatory practices and fitness-for-service guidelines in Sweden . . . . . 128
7.2.15. Fitness-for-service guidelines in Switzerland . . . . . . . . . . . . . . . . . . . . . . 129
7.3. Statistical methods for degradation growth assessment . . . . . . . . . . . . . . . . . . . . . 129
7.4. Use of probabilistic fracture mechanics analysis to assess
steam generator ageing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
8. STEAM GENERATOR MAINTENANCE: MITIGATION, REPAIR

AND REPLACEMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
8.1. Primary-side mitigation techniques for tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

8.1.1. Rotopeening and shot peening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
8.1.2. Stress relieving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
8.1.3. Reducing the hot-leg side temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
8.2. Secondary-side mitigation techniques for tubes . . . . . . . . . . . . . . . . . . . . . . . . . . 137
8.3. Tube repair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
8.3.1. Plugging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
8.3.2. Sleeving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
8.3.3. Nickel plating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
8.4. Vibration control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
8.4.1. Preheater repairs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
8.4.2. Antivibration bar replacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
8.5. Mitigation of the thermal fatigue of the feedwater nozzles and piping . . . . . . . . . 147
8.6. Repair of feedwater nozzles and piping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
8.7. Steam generator replacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
8.7.1. PWR replacement steam generator designs . . . . . . . . . . . . . . . . . . . . . . . 150
8.7.2. CANDU replacement designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
8.7.3. WWER replacement designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
9. STEAM GENERATOR AGEING MANAGEMENT PROGRAMME . . . . . . . . . . . . . 153
9.1. Understanding steam generator ageing and feedback of operating experience ... 155
9.2. Definition of steam generator AMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
9.3. Steam generator operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
9.4. Inspection, monitoring and assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
9.4.1. Inspection and monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
9.4.2. Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
9.5. Maintenance: mitigation, repair and replacement . . . . . . . . . . . . . . . . . . . . . . . . . 158
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
ABBREVIATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
CONTRIBUTORS TO DRAFTING AND REVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

i.
1.1. BACKGROUND
Managing the safety aspects of nuclear power plant (NPP) ageing requires
implementation of effective programmes for the timely detection and mitigation of ageing
degradation of plant systems, structures and components (SSCs) important to safety, so as to
ensure their integrity and functional capability throughout plant service life. General guidance
on NPP activities relevant to the management of ageing (maintenance, testing, examination
and inspection of SSCs) is given in the IAEA Nuclear Safety Standards (NUSS) Code on the
Safety of Nuclear Power Plants: Operation (Safety Series No. 50-C-O, Rev. 1) and associated
Safety Guides on in-service inspection (50-SG-O2), maintenance (50-SG-O7, Rev.l) and
surveillance (50-SG-O8, Rev.l).
Trie Operation Code requires that NPP operating organizations prepare and carry out a
programme of periodic maintenance, testing, examination and inspection of plant systems,
structures and components important to safety to ensure that their level of reliability and
effectiveness remains in accord with the design assumptions and intent and that the safety
status of the plant has not been adversely affected since the commencement of operation. This
programme is to take into account the operational limits and conditions, any other applicable
regulatory requirements, and be re-evaluated in the light of operating experience. The
associated Safety Guides provide further guidance on NPP programmes and activities that
contribute to timely detection and mitigation of ageing degradation of SSCs important to
safety.
The Safety Guide on In-Service Inspection (ISI) provides recommendations on methods,

frequency and administrative measures for the ISI programme for critical systems and
components of the primary reactor coolant system aimed at detecting possible deterioration
due to the influences of stress, temperature, irradiation, etc. and at determining whether they
are acceptable for continued safe operation of the plant or whether remedial measures are
needed. Organizational and procedural aspects of establishing and implementing an NPP
programme of preventive and remedial maintenance to achieve design performance throughout
the operational life of the plant are covered in the Maintenance Safety Guide. Guidance and
recommendations on surveillance activities, for SSCs important to safety, (i.e. monitoring
plant parameters and systems status, checking and calibrating instrumentation, testing and
inspecting SSCs, and evaluating results of these activities) are provided in the Surveillance
Safety Guide. The aim of the surveillance activities is to verify that the plant is operated
within the prescribed operational limits and conditions, to detect in time any deterioration of
SSCs as well as any adverse trend that could lead to an unsafe condition, and to supply data
to be used for assessing the residual life of SSCs. The above Safety Guides provide general
programmatic guidance, but do not give detailed technical advice for particular SSCs.
Ageing management specific programmatic guidance is given in Technical Reports Series

No. 338 "Methodology for the Management of Ageing of Nuclear Power Plant Components
Important to Safety" and in a Safety Practice No. 50-P-3 "Data Collection and Record
Keeping for the Management of Nuclear Power Plant Ageing". Guidance provided in these
reports served as a basis for the development of component specific technical documents
(TECDOCs) on the Assessment and Management of Ageing of Major NPP Components
Important to Safety. This publication on Steam Generators is one of such TECDOCs.
The steam generators in the pressurized water reactor (PWR), Canada deuterium-
uranium (CANDU) reactor, and Russian water moderated, water cooled energy reactor
(WWER) plants are large heat exchangers that use the heat from the primary reactor coolant
to make steam in the secondary side to drive turbine generators. A typical plant has two to
six steam generators per reactor; although some units have up to twelve steam generators. The
steam generators are shell-and-tube heat exchangers each with several thousands of tubes. The
primary reactor coolant passes through the tubes and boils water on the outside of the tubes
(secondary side) to make steam. The design confines radioactivity from neutron activation
or fission products to the primary coolant during normal operation. However, the primary
reactor coolant is at a higher pressure than the secondary coolant, so any leakage from defects
in the tubes or in the WWER collectors is from the primary to the secondary side, and rupture
of the heat exchanger tubing or the WWER collectors can result in release of radioactivity
to the environment outside the reactor containment through the pressure relief valves in the
secondary system.
The thin-walled steam generator tubes are, therefore, an important part of the reactor
coolant pressure boundary and, in fact, can comprise well over 50% of the area of the total
primary system pressure-retaining boundary. To act as an effective barrier, this tubing must
be essentially free of cracks, perforations, and general deterioration. However, widespread
degradation of the steam generator tubes has occurred at a number of plants. As a result,
many steam generator tubes have been removed from service by plugging or repaired with
sleeves. Other tubes with small defects remain in service.
Certain accidents, such as a sudden break in the steam line, can lead to rapid
depressurization of the secondary coolant system. The pressure difference across the tubing
walls generated during these accidents may result in simultaneous leakage or rupture of a
number of steam generator tubes or rupture of a WWER collector when an active degradation
mechanism has damaged a large number of tubes or the collector. Simultaneous leakage or
rupture of several tubes or a WWER collector can lead to a plant transient which is difficult
to control and radioactivity levels released to the environment which may exceed site limits.
The sudden rupture of several steam generator tubes or a WWER collector also results in a
rapid depressurization of the primary coolant system and possibly may uncover the core and
cause core melting.
Steam generator performance is important to nuclear power plant safety. For example,
the various nuclear power plants in the United States of America have a core damage
frequency which ranges from a low of 1 * 10"6 per year to about 5 x 10"4 per year. Steam
generator tube rupture accidents are relatively small contributors to these values, but are risk
significant due to containment bypass. A review of 20 US PWR Individual Plant
Examinations (IPEs) has shown that the risk associated with steam generator tube ruptures
can be as high as 75% of the total plant risk. These risk significant accidents can be induced
by operational transients and rare events with degraded steam generator tubes which could
lead to core rnelt (Ellison et al. 1995).
To summarize, the major safety function of the steam generator is to act as a barrier
between the radioactive primary side and the non-radioactive secondary side. Any degradation
mechanism which inpairs this function, i.e. which may lead to either a single or multiple tube
rupture or to simultaneous failure (leakage) of several tubes under certain accident conditions
or failure of a WWER collector, is a significant safety concern. There is always some risk
associated with the operation of steam generators with respect to the safety concerns
mentioned above. This TECDOC describes the various approaches, practices and requirements
used in different countries to obtain a risk of acceptable level.
1.2. OBJECTIVE
The objective of this TECDOC is to document the current practices for the assessment
and management of the ageing of nuclear power plant steam generators. The TECDOC
emphasizes safety aspects and also provides information on current inspection, monitoring and
maintenance practices for managing ageing of steam generators.
The underlying objective of this TECDOC series is to ensure that the information on the
current assessment methods and ageing management techniques is available to all involved,
directly and indirectly, in the operation of nuclear power plants in the IAEA Member States.
The target audience includes nuclear power plant operators, regulators, technical support
organizations, designers, and manufacturers.
1.3. SCOPE
This TECDOC documents current practices for the assessment and management of
ageing of the following types of steam generators used in water cooled nuclear power plants:
(a) verticalAubesheet boiling steam generators, commonly known as "recirculating vertical U-
tube steam generators;" (b) vertical/tubesheet super heated steam generators, commonly
known as "once-through steam generators," and (c) horizontal/collector boiling steam
generators used in WWER reactors.
The steam generator subcomponents discussed in this publication are those susceptible
to ageing damage and whose consequence of failure has a significant safety impact as
discussed in Section 1.1; the steam generator tubes, tubesheets or collectors, plugs (tube and
tubesheet), and sleeves (i.e. components whose failure impairs the primary to secondary
pressure boundary). In addition, this TECDOC also discusses two other subcomponents:
feedwater nozzles and shell girth welds. These components have experienced significant
degradation in some plants but their failure is a secondary side pressure boundary failure
(affects conventional safety) and does not immediately lead to any release of radioactivity.
The TECDOC does not address life or life-cycle management of steam generators because
it is written from the safety perspective and life management includes economic planning.
1.4. STRUCTURE
The steam generator designs are discussed in Section 2. The design bases for the
components of interest to steam generator ageing are presented in Section 3. The stressors,
susceptible sites, and failure modes associated with the various steam generator degradation
mechanisms are presented in Section 4. These degradation mechanisms include primary water,
outside diameter, and transgranular stress corrosion cracking; fretting, wear, and thinning;
pitting; denting; high-cycle fatigue; wastage; erosion-corrosion; and corrosion-fatigue.
Operational guidelines aimed at preventing or minimizing age related degradation of steam
generators are discussed in Section 5. Tubing inspection requirements and technologies are
discussed in Section 6. Fitness-for-service guidelines in various countries are presented in
Section 7. Mitigation, repair, and replacement technologies are discussed in Section 8. The
TECDOC concludes, in Section 9, with guidelines for a systematic steam generator ageing
management programme.
2. STEAM GENERATOR DESCRIPTIONS
This section describes the different designs for currently operating steam generators.
Recirculating steam generators (RSGs), designed by Westinghouse (USA), Combustion
Engineering (USA), Framatome (France), Mitsubishi Heavy Industries (Japan), and
Siemens-Krafrwerke Union (Germany), are described first. The Canadian designs are
discussed next with the Babcock & Wilcox (USA) once-through steam generator design and
the Russian (WWER) designs completing the section. Emphasis is placed on the design
aspects and fabrication methods which may affect steam generator degradation.
2.1. PWR REORCULATiNG STEAM GENERATORS
hi RSG, the primary system coolant flows through U-tubes with a tubesheet at the
bottom of the generator and U-bends at the top of the tube bundle (Fig. 1). Primary coolant
enters the steam generator at 315-330°C on the hot-leg side and leaves at about 288°C on the
cold-leg side. The secondary system water (feedwater) is fed through a feedwater nozzle, to
a feedring, into the downcomer, where it mixes with recirculating water draining from the
moisture separators. This downcomer water flows to the bottom of the steam generator, across
Steam outlet
Upper
tubesheet
Auxiliary
feedwater
inlet Steam
separators
Steam Feedwater
outlet inlet
K<
Downcomer
Feedwater annul us
inlet
Feed
annuhis
Aspirating
steam
Tube
sheet
Lower
tube
sheet
Primary inlet Primary outlet
Once-through Recirculating
steam generator steam generator
FIG. 1. PWR steam generator cross sections (EPRI1985). Copyright Electric Power
Research Institute; reprinted with permission (left).
Main steam _^™nnx^ ^^v^ Dryandsaturated
steam leaves vessel
Steam/water mixture
Water level
Auxiliary feedwater
nozzle
Steam/water mixture
leaves tube bundle
Preheater '
l____J
/Main feedwater nozzle

X^
Primary fluid inlet Primary fluid outlet
FIG. 2. Typicd design of a steam generator with apreheater (CSGORG 1983).
the top of the tubesheet, and then up through the tube bundle where steam is generated About
25% of the secondary coolant is converted to steam on each pass through the generator, the
remainder is recirculatecL
Some RSGs include economizer sections (preheaters), which are separate sections in the
steam generator near the cold leg outlet, shown in Fig. 2. The feedwater flows into the
preheater through a nozzle located in the lower part of the vessel and there is no feedring in
these steam generators. Auxiliary feedwater is injected through a separate nozzle in the upper
part of the vessel. Heat from the primary fluid leaving the steam generator is used to preheat
the feedwater to near the saturation temperature before it is mixed with the recirculating
secondary system coolant.
TABLE I. TYPICAL US STEAM GENERATOR MODELS AND THEIR PARAMETERS
Manufacturer type
and model Wesnnghouse (recirculating)
24 27 33 44' 51 A-M"* D2/D3' D4'
Heat transfer area 24834 27 700 33340 44500 51 500 48000 47 000
(ft2)h
No of tubes 2604 3794 2604 3260 3388 4674 4578
No of row- 1 tubes 82 100 82 92 94 114 114
Tube pattern Square Square Square Square Square Square Square
Tube spacing (in f 1 2187 1 026 or 1 031 1 25 1 200 or 1 234 1 281 1 063 1 063
Tube dimensions 0 875x0 050 0 750x0 055 0 875x0 050 0 875x0 050 0 875x0 050 0 750x0 043 0 750x0 043
(in)
Tubing material Alloy 600 Alloy 600 Alloy 600 Alloy 600 Alloy 600 Alloy 600 Alloy 600
Tubing heat Mill-annealed Mill-annealed Mill-annealed Mill-annealed Mill-annealed Mill-annealed Mill-annealed

treatment
Tubesheet Part-depth Part-depth Part-depth Part-depth Pan-depth Full-depth Full-depth

expansion method rolled rolled rolled rolled rolled1 rolled rolled
Tubesheet crevice 1825 18 18 18 19 or 20 18 18 75 or None None

depth (in )c 19"
Tube support type Drilled hole Drilled hole Drilled hole Drilled hole Drilled hole Drilled hole Drilled hole
Tube support Carbon steel Carbon steel Carbon steel Carbon steel Carbon steelf Carbon steel Carbon steel
material
Preheater type None None None None None Split flow Counterflow
expanded
preheater
tubes
Flow distribution None None None None None' D2 no Yes

baffles D3 yes
' Replacement Models 44F 51F and 54F use hydraulically expanded, thermally treated Alloy 600 tubing and 405 stainless steel tube support plates except for
the model 54Fs at D C Cook and Indian Point Unit 3 which have thermally treated Alloy 690 tubing The replacement models generally match the heat transfer
area of the steam generators they replaced except for the 54 Fs with Alloy 690 tubing which are slightly larger than the original 51s due to the slightly lower
thermal heat transfer properties at the Alloy 690 material vis-a-vis the Alloy 600 material
h
1 fr = 0 093 nr 1 in = 25 4 mm
' Later Model 51s used full-depth rolled or explosively expanded tubes The tubesheet thickness ranges from 525 to 610 mm
Table I lists the design features for eleven Westinghouse and two Combustion
Engineering type steam generator models. Table II lists the design features of seven
Mitsubishi Heavy Industries steam generator models and Table HI lists the design features
of the steam generators delivered by Siemens/KWU.
2.2. CANDU RECIRCULATING STEAM GENERATORS
Currently operating CANDU steam generators are vertical RSGs built by Babcock & Wilcox
Canada Ltd. The only exception is the Wolsung 1 unit in the Republic of Korea which uses
TABLE I. (cont.)
Manufacturer type B&W once- Combustion Engineering

and mode] Wesunghouse (recirculaung) through (recirculating)
D5 E[ F A75 177 67 80
Heat transfer area 47000 50000 50000 75 180 132500 90700 N/A
(ft1)"
No of tubes 4570 4864 5626 6307 15 531 8519 11 012
No of row -1 114 120 122 70 - 167 N/A

tubes
Tube pattern Square w/T slot Square w/T slot Square w/T slot Triangle Triangle Triangle Triangle
Tube spacing (in )' 1 063 1 080 0980 0980 0875 0 974 1 00 1 000
Tube dimensions 0 750 X 0 043 0 750 x 0 043 0 688 x 0 040 0688 x 0 625 X 0 034 0 750 X 0 048 0 750 x 0 042
On) 0040
Tubing heat Thermally Mill-annealed or Thermally Thermally Mill-annealed Mill-annealed Mill-annealed

treatment treated therm treated treated treated
Tubesheet Hydraulic Full-depth rolled Hydraulic Hydraulic Partial -depth Explosive Explosive
expansion method or hydraulic rolled
Tubesheet crevice None None None None 22 None None

depth (in )'
Tube support type Broached Drilled Broached Broached Broached Eggcrate/ Eggcrate/
quatrefoil quatrefoil trefoil trefoil vertical vertical
Tube support Stainless steel Carbon or 405 stainless 405 stainless Carbon or Carbon steel Stainless steel
material stainless steel steel MnMo steel
Preheater type Counterflow, Counterflow, None None None None Axial flow
expanded pre- expanded
heater tubes preheater tubes
Flow distribution Yes Yes Yes Broached No None Yes

baffles plate
d
For Model 5K with part-depth rolled tubes only
The crevice radial gaps varied from 0005 to 0011 inches except in the Model ?4 where they were 00135-00175 inches
'Some later Model 51 s were equipped with alloy steel tube support plates and flow distribution baffles
The row 1 and 2 tubes in most Model 51 D2/D3, D4 and E steam generators have been u-bend heat treated and shot or rotopeened for added resistance to
PWSCC
similar steam generators built by Foster Wheeler. Atomic Energy of Canada Limited, and for
some units Ontario Hydro, selected the key design parameters for the CANDU steam
generators including the tubing material and size, the steam generator size, and the key
thermal hydraulic parameters. The fabricators did the detailed design of the equipment.
CANDU RSGs are very similar to the PWR RSG with some subtle differences in size,
materials, operating temperatures and tube support structure. Figure 3 depicts the steam
generator used in the Darlington Generating Station which has all the most current features
of CANDU RSGs.
8
TABLE II. TYPICAL MITSUBISHI HEAVY INDUSTRIES RECIRCULATING STEAM GENERATOR MODELS AND THEIR PARAMETERS
Manufacturer MHI MHI MHI MHI MHI MHI MHI

and Model 44 46F 51.51A 51M 51F.51FA 52F,52FA 54F.54FA
Heat transfer area (m2) 4130 4300 4785 4780 4780 4870 5055
No. of tubes 3260 3382 3388 3382 3382 3382 3382
No. of row- 1 tubes 92 94 94 94 94 94 94
Tube pattern Square Square Square Square Square Square Square
Tube spacing (mm) 31.35 32.54 32.54 32.54 32.54 32.54 32.54
Tube dimensions (mm) 22.23xl.27 22.23xl.27 22.23xl.27 22.23xl.27 22.23xl.27 22.23xl.27 22.23xl.27
Tubing heat treatment Mill-annealed Thermally treated Mill-annealed Mill-annealed, Thermally treated Thermally treated Thermally treated
thermally treated
Fubcsheet expansion Part-depth Full-depth hydraulic Part-depth rolled, Full-depth rolled, Full-depth hydraulic Full-depth hydraulic Full-depth hydraulic
method rolled and one step rolled full-depth rolled full-depth hydraulic and one step rolled and one step rollec
and rolled and one step rolled
lubesheet crevice 497' None 488', None None None None None
depth (mm) (original design'
Tube support type Drilled Broached eggcrate Drilled Drilled, Broached eggcrate Broached eggcrate Broached eggcrate
drilled chamfer
Tube support material Carbon Steel 405 stainless steel Carbon steel Carbon steel, 405 stainless steel 405 stainless steel 405 stainless steel
405 stainless steel
Preheater type None None None None None None None
Flow distribution None Yes None Yes Yes Yes Yes

baffles
"Tubesheet radial gap of 0.185 mm

TABLb III TYPICAL SIEMENS/KWU REURCULATING STEAM GENERATORS AND THEIR PARAMETERS
Manufacturer MAN-GHH MAN-GHH Balcke Babcock " Standatd with preheater MAN-GHH * Replacement SGs for
and model Obrigheim(Orig.) Obrighelm (Repl.) Slade BiMis A Konvoi '» 51O51M/D3
Heat transfer area (m?) 2750 3070 2930 4510 5386 5427 5105/6103/7155 '»
No of tubes 2605 3010 2993 4060 4086 4118 51 30/5428 m)
No of row- 1 tubes 81 46 49 55 48 54 57/59 m)
Tube pattern Rectangular Triangular Triangular Tnangular Tnangular Triangular Triangular
Tube spacing (mm) 279x288 290 293 300 300 300 26164
Tube dimensions (mm) 22 x 1 23 (1 5) '> 22 x 1 23 22 x 1 23 22 x 1 23 22 x 1 23 22 x 1 23 1905 x 109
Tubing material Alloy 600 Alloy 800 M p> Alloy 800 M p) Alloy 800 M p) Alloy 800 M "> Alloy 800 M p) Alloy 690 n)
Alloy 800 M"
Tubing heat treatment Mill annealed s> 5) !) g) !) Alloy 690 therm, treated
Alloy 800 M"^
Tubesheet expansion method 'art-depth rolled 'art-depth rolled Part-depth rolled Part-depth rolled Part-depth rolled Part-depth rolled Full-depth hydraulic plus
3 locations) (both ends) (both ends) (both ends) (both ends) (both ends) sart-depth rolled (both
ends)
TS crevice depth (mm) None *Jone None None None None None
Tube support type Eggcrate b) Bggcrate c) Eggcrate d> Eggcrate e) iggcrate c) Eggcrate c) Eggcrate °
Tube support material Stainless steel Stainless steel Stainless steel Stainless steel Stainless steel Stainless steel Stainless steel
Prcheater type None None None None Split-flow design None None
Flow distribution baffles None Yes None None Yes None Yes
U-Bend Treatment None None None None None None Alloy 690 Yes°>
Peenmg of the roll-transition None None None None None None None
/one
Notes
a) Innermost rows wall thickness = I 5 mm
b) Bend Vertical flat bars k) Replacement steam generator for Westmghouse model 51C (Rmghals 2, Manuf
c) Bend Vertical and horizontal flat bars, vertical corrugated strips MAN-GHH), 51M (Doel 3, Manuf ENSA/CMI), D3 (Asco 'A and Almaraz 'A, Manuf
d) Bend Vertical flat bars, horizontal and vertical corrugated strips ENSA), D3 (Rmghals 3, Manuf Framatome)
c) Bend Radial flat bars, vertical corrugated strips I) Replacement steam generator for Rmghals 2 5105m 2 , Doel 3, Asco 'A and Alrmraz Vi
0 Bend Vertical and horizontal flat bars, vertical corrugated ships, block tubmg (zero gaps) 6103m2, Ringhals3 7155m 2
g) Similar to ASTM SB 163, UNS N05800 m) Replacement steam generator for Rmghals 2, Doel 3, Asco 'A and Almaraz 'A 5130 tubes, 57
h) Grafenrhemfeld, Grohnde (Manuf MAN-GHH), Brokdorf (Manuf UDDCOMB), row-1 tubes, Rmghals 3 5428 tubes, 59 row-1 tubes
Tnllo 1 (Manuf ENSA) n) Replacement generator for Rmghals 2 and Rmghals 3 Alloy 690, Doel 3, Asco !/2 and
j) Almost identical plants Isar 2, Neckarwestheim 2, Emsland Almaraz'/2 Alloy 800 M
o) Alloy 690 Tubes with Radius <300 mm
p) Modified according to Siemens/KWU specification
Legend
1. Steam drum
2. Steam generator
3. Heavy water inlet (2)
4. Heavy water outlet
5. Downcommerannulus
6. Primary cyclones
7. Secondary cyclones
8. Slowdown nozzles
9. Manway
10. Main steam outlet nozzles
11. Preheater
12. Steam generator support
13. Contaminant seal bar/skirt
14. Lower lateral restraint lugs
15. Reheat condensate return nozzle
16. Feedwater nozzle
17. Upper lateral restraint lugs
18. Ring beam
FIG. 3. CANDU Recirculating Steam Generator used at the Dariington station. This design
is typical of the current CANDU models. (Courtesy of C. Mantska, Ontario Hydro.)
Although the size of CANDU RSGs has escalated greatly with successive reactor
designs, they are generally smaller than PWR RSGs, and operate at lower temperatures
(290°C to 310°C primary inlet temperature). The lower temperatures generally delay the onset
of thermally activated corrosion processes such as primary water stress corrosion cracking
(PWSCQ or intergranular stress corrosion cracking (IGSCQ. Because the primary coolant
in a CANDU reactor is heavy water (D2O), relatively small tube sizes [12.7 mm 0/2") OD
and, in more recent units, 15.9 mm (5/8") OD] have been used to minimize the heavy water
inventory. The smaller size of the primary (lower) head and tubes increases the difficulty in
performing maintenance activities such as tube inspection, plugging, removal, etc. The
nominal tube wall thickness ranges from 1.13 mm to 1.2 mm depending on the type of tube
alloy used (for example Alloy 800M has a lower thermal conductivity than Alloy 600
requiring thinner tubes).
The most important area of diversity in the CANDU design is in the choice of tube
material, the CANDU steam generators currently operate with tubes made from high
temperature, mill annealed Alloy 600, Monel 400 and titanium stabilized Alloy 800. These
materials are susceptible to different types of degradation.
11
1. Steam generator pressure vessel 4. Moisture separator
2. Primary collector 5. Steam-collector
3. Heat exchange tubes 6. Feedwater inlet
FIG. 4a WWER-440 steam generator (side view). (Courtesy of Y. G. Dragunov, OKB

Gidropress.)
Steam pipe
Hot
collector
Tubing
Inlet
FIG. 4b. WWER-440 steam generator (end view). (Courtesy of Y. G. Dragunov, OKB
Gidropress.)
12
10
FIG. 5a Cut-away drawing of a WWER-1000 steam FIG. 5b. WWER-1000 steam generator (end view) (Titov
generator (Koryakin 1993). Key: 1-Steam generator drum; 1991). Key: 1-SG shell; 2-Tube bundle; 3-Feedwater branch
2-Cold header; 3-Hot header; 4-Heat exchanger tubes; 5- pipe; 4-Separation device; 5-Steam collecting heater; 6-Point
Submerged perforated separator; 6-Feedwater header; 7- of header jamming; 7-Immersed perforated sheet; 8-
Steani separators (Koryakin 1993). Copyright Nuclear Unperforated section in perforated zone; 9-Inlet ("hot")
Engineering International; used with permission. header; 10-Outlet ("cold") header. Courtesy of Y. G.
Dragunov, OKB Gidropress. Copyright Nuclear Engineering
International; used with permission.
2.3. PWR ONCE-THROUGH STEAM GENERATORS
US once-through steam generators use straight heat exchanger tubes with tubesheets at both
the top and bottom of the steam generator, as shown in Fig. 1. Primary coolant is pumped
through the tubes from top to bottom while the secondary coolant moves around the outside
of the tubes from bottom to top in a counter-flow direction. The secondary-system water
enters a feed annulus above the ninth tube support plate level where it mixes with steam
aspirated from the tube bundle area and is preheated to saturation. The saturated water flows
down the annulus, across the lower tubesheet, and up into the tube bundle where it becomes
steam. This superheated steam flows radially outward and then down the annulus to the steam
outlet connection. Most of the secondary coolant is completely evaporated in a single pass
through the steam generator.
FIG. 6a Top view sketch of the tube FIG. 6b. Basic arrangement of the heat
layout in WWER-440 and WWER-1000 exchanger tubes and headers used in the
steam generators. (Courtesy of Y. G. WWER-1000 steam generators (Titov 1991).
Dragunov, OKB Gidropress.) Copyright Nuclear Engineering International,
reprinted with permission.
14
TABLE IV. WWER STEAM GENERATOR PARAMETERS
Parameters WWER-440 WWER-1000 WWER-

1000U
Thermal power, MW 229.2 750 750
Steam capacity, kg/s 125 408.33 408.33
Pressure of steam, MPa 4.61 6.27 6.27
Steam temperature, °C 258.9 278.5 278.5
Feedwater temperature, °C 164-223 164-220 164-220
Coolant temperature, °C
- at steam generator inlet 295 320 320
- at steam generator outlet 267 290 292
Coolant flow rate, rrrVhr 7100 21 200 21 200
Coolant pressure, MPa 12.26 15.7 15.7
Coolant flow rate in tubes, m/s 2.71 4.21 4.91
Average heat transfer factor, kW/nfK 4.7 5.4 6.1
Mean logarithmic temperature head, °C 18.7 22.9 24
Specific heat flux (average), kW/m2 89.23 123 141
Total heat exchanging surface, m2 2576.6 6115 5126.6
Total number of tubes 5536 11000 9157
Diameter and thickness of tube walls, mm 16x1.4 16x1.5 16x1.5
Tube mean length, m 9.26 11.10 11.14
Pressure loss along the coolant path, MPa 0.075 0.126 0.169
Reduced outlet steam rate from the evaporation surface, m/s 0.240 0.382 0.382
Steam humidity at steam generator outlet, % 0.25 0.2 0.2
Vessel material 22K 10GN2MFA 10GN2MFA
Collector material 08X18N10P 10GN2MFA 08X18N10T
with inner cladding perforated area
Heat exchanging tube material 08X18N10T 08X18N10T 08X18N10T

Collector dimensions in the perforated area
- inner diameter, mm 800 834 780
- wall thickness, mm 136 171 198
with cladded layer
Hole array in the header perforated area

- dimensions of minimum ligament, mm 11.34 6.93b 9.75
- number of horizontal rows along the height 77 110 94
- number of tubes in a horizontal row 89 120 112
15
TABLE IV. (Coot.)
Parameters WWER-440 WWER-1000 WWER-

1000U
square array staggered staggered
24 19 22.1
Tube array in tube bundle 29.5 23 25, 23C
- array pitch along the vertical axis, mm
- array pitch along the horizontal axis, mm
Submerged perforated sheet absent installed installed
Steam generator circulating factor (minimum) 4-6 1.5 1.9
Void fraction, % 0.32 0.493 0.485
'This material is also labelled OSCrlSNilOTi which is a titanium stabilized austenitic stainless steel with 0 08% carbon, 18% chrome,
10% nickel and less than 1% titanium
h
Along medium surface
'25 mm for the central set and 23 mm for the lateral set
2.4. WWER STEAM GENERATORS
The steam generators used in the Russian designed WWER-440 and WWER-1000
plants are horizontal shell-and-tube heat exchangers manufactured by ZiO (Podolsk, Moscow
Region), Atommash (Volgodonsk, Volgograd Region) and Vitkovice (Czech Republic). They
consist of a pressure vessel, a horizontal heat exchange tube bundle, two vertical primary
collectors, a feedwater piping system, moisture separators and steam collector. A sketch of
a WWER-440 steam generator is shown in Figs 4a and 4b (side and end views). A sketch of
a WWER-1000 steam generator is shown in Figs 5a and 5b. The tube bundle arrangement in
the WWER-440 and WWER-1000 steam generators, as seen from the top, is shown in Fig. 6.
Primary coolant enters the steam generator through a vertical collector, travels through the
horizontal U-shaped submerged stainless steel tubing, and exits through a second vertical
collector. The tube ends penetrate the collector wall (which performs the same function as the
tubesheet in a PWR steam generator) and are expanded using either a hydraulic or explosive
expansion process and then welded at the collector inside wall surface. The WWER-440
collectors are made of Ti-stabilized austenitic stainless steel. The WWER-1000 collectors are
made of low-alloy steel with higher tensile properties, clad with stainless steel. The WWER-
440 tubes are arranged in line (corridor). The WWER-1000 tubes are staggered Grids
consisting of stainless steel bars and stamped wave-like plates are used to separate and
support the tubes. The distance between the tube supports is 700-750 mm.
16
The steam generator vessel is a carbon steel (WWER-440) or low-alloy bainitic steel
(WWER-1000) horizontal cylinder consisting of forged shells, stamped elliptical ends and
stamped branch pipes and hatches welded together as shown in Figs 4 and 5. The vertical hot
and cold primary coolant collectors penetrate the vessel near its mid-point. Feedwater is
supplied to the middle of the WWER-400 tube bundle by perforated piping. In the WWER-
1000 steam generators, the feedwater is supplied to the top of the hot side of the tube bundle
under a submerged perforated sheet. The tube bundle is completely submerged in both
designs.
The WWER-440 and WWER-1000 steam generator designs are similar except for the
(a) size (the WWER-1000 steam generator is about 4 meters longer), (b) tube arrangement
(corridor versus staggered), (c) collector material, (d) feedwater supply location, (e)
submerged perforated top plate (WWER-1000 only), (f) steam dryer arrangement, (g)
emergency feedwater distribution system (WWER-1000 only), (h) steam header arrangement,
(I) and vessel material. The WWER-IOOOU steam generator has been designed to replace the
original WWER-1000 steam generators as needed The WWER-IOOOU has the perforated
region of the collectors fabricated from austenitic stainless steel. Table IV lists the WWER-
440, WWER-1000, and WWER-IOOOU design features.
17
BASK,
FABRICATION AND MATERIALS
This section mentions briefly the code requirements used to design steam generators and
discusses some of the more important steam generator fabrication practices and materials of
construction. Most of the information presented in this section concerns the heat exchanger
tubing, including its fabrication, materials of construction and installation into the tube sheets
(or collectors). The tube support, feedwater nozzle and steam generator shell designs and
materials are also discussed. This section builds on the design and material information
presented in Section 2 and concentrates on specific sites susceptible to the degradation
mechanisms described in Section 4.
3.1. CODES AND SPECMCATIONS
Although many countries have, or are developing their own standards and codes for the
design of nuclear power plant components, the load restrictions are generally based on Section
IQ of the American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel
Code. The objective of designing and performing a stress analysis with the rules of Section
HI is to afford protection of life and property against ductile and brittle failure. The ASME
Class 1 design requirements are used for all the primary side pressure retaining components.
The components on the secondary side are required to satisfy ASME Class 2 requirements.
However, common practice is to design the entire steam generator shell to the ASME Class
1 requirements. Therefore, Article NB-2300 of Section IE of the ASME Code is referred to
for assurance of adequate fracture toughness of all pressure retaining materials in the steam
generator. In addition, the steam generator tube/tubesheet complex meets the stress limitations
and fatigue criteria specified in the ASME Code. The requirements of Section HI are
discussed in detail in a companion publication in the IAEA TECDOC series entitled
Assessment and Management of Ageing of Major Nuclear Power Plant Components Important
to Safety: Pressurized Water Reactor Pressure Vessels. The NPP design requirements in
Germany and Russia, which differ from the ASME requirements, are also discussed in that
document.
3.2. FABRICATION AND MATERIALS
Materials and methods used to fabricate steam generator components significantly affect
their susceptibility to corrosion, especially to stress corrosion cracking. Degradation of the
steam generator tubing is also influenced by other aspects of the steam generator design and
construction, such as the tube support design and the method of tube installation.
3.2.1. Heat Exchanger Tubes
Initially, the heat exchanger tubing in most of the PWR steam generators placed in-
service in the western countries (except Germany) was made from nickel based Alloy 600
(76% Ni, 15.5% Cr, 8% Fe, < 0.15% Q. The German steam generators designed by
Siemens/KWU use Alloy-800M tubing. Now, most steam generators designed by
Westinghouse, Framatome, Siemens/Framatome, Babcock & Wilcox and Mitsubishi-Heavy
Industries are being fabricated with thermally treated Alloy 690 (61% Ni, 29.5% Cr, 9% Fe,
18
<Q.025% Q. Siemens/KWU and Babcock & Wilcox Canada are also supplying replacement
steam generators with Alloy-800M tubing.
Tube fabrication generally starts with extrusion of a shell from an ingot and then several
cold reduction steps by either drawing or pilgering. Each reduction step is followed by mill-
annealing, which typically consists of passing tube lengths through a furnace on a traveling
belt at temperatures high enough to recrystallize the material and dissolve all the carbon
(about 980°C or above).
The mill-annealing temperature and initial carbon content are two of the important
parameters in controlling the mechanical and corrosion behaviour of nickel based alloys such
as Alloy 600. The object of the mill-annealing steps is to first dissolve all the carbides and
obtain a relatively large grain size and then cover the grain boundaries with carbides upon
slow cooling in air. A higher carbon content requires a higher mill-annealing temperature to
dissolve all the carbides. Undissolved intragranular carbides are undesirable because they
provide nucleation sites for the dissolved carbides and prevent precipitation of the carbides
on the grain boundaries and, therefore, prevent appropriate grain boundary carbide coverage.
The mill-annealing temperature also controls the material yield strength and, therefore, the
residual stresses. Higher mill-annealing temperatures result in lower residual stresses (in tubes
which are not stress relieved). Starting in the late 1970s, the mill-annealed Alloy 600 tubes
from some vendors were also given a final thermal treatment at about 705°C for 15 hours in
order to relieve fabrication stresses and to further improve the microstructure. The thermal
treatment process promotes carbide precipitation at the grain boundaries and diffusion of
chromium to the grain boundaries. Therefore, somewhat lower mill-annealing temperatures
can be used with the same final grain boundary carbide coverage and the chromium used to
form the chromium carbides is replenished on the grain boundary. Alloy 600 tubing with
grain boundary chrome depletion is susceptible to outside diameter stress corrosion cracking
(ODSCQ. Alloy 600 tubing with insufficient carbides on the grain boundaries is susceptible
to PWSCC.
Subsequent to the final mill-anneal, the tubing is passed through roll straighteners to
produce a straight product. The straightening process plastically deforms the tubing, imparting
some residual stresses. After straightening, the tubing may be abrasively polished (e.g., using
belt abrasives) to remove about 0.025 mm from title exterior surface. This step removes
surface imperfections, but also results in the tubes having a thin cold-worked surface layer
and significant residual surface stresses, which can range from compressive to highly tensile.
The final manufacturing steps for straight tubes involve visual, ultrasonic, and
eddy-current inspections as well as various cleaning operations, including blasting the interior
surfaces with ceramic grit. For RSGs, the straight tubes are bent to the desired U-tube
configuration. For tight radius bends, internal mandrels are often used to minimize ovality of
the bent portion of the tube [Shah et al. 1992]. In addition, the tight radius U-bends of tubes
in some of the existing steam generators which had not been thermally treated, were stress
relieved at 705°C for at least 5 minutes to relieve bend-induced stresses.
The annealing and thermal treatment temperatures and other details of the tube
processing were somewhat different for the various manufactures and steam generator models
and are briefly discussed below.
19
Babcock & Wilcox Practice. Babcock & Wilcox practice was to mill-anneal at a
relatively high temperature, about 1065-1095°C [Jones 1982]. In addition, after tube
installation, Babcock & Wilcox heat treated the entire steam generator at about 595°C for 15
hours to reduce residual stresses from tube fabrication and installation (e.g., at roll
transitions), and to increase resistance to PWSCC by developing carbides at grain boundaries.
However, it also resulted in sensitization (chromium depletion at grain boundaries), making
the tubing susceptible to other forms of corrosion (stress corrosion cracking in oxidizing
acidic conditions).
Combustion Engineering Practice. The Combustion Engineering tubing was annealed at

a relatively high temperature of 980-1065°C [Owens 1987a]. This final mill-anneal resulted
in relatively large grain sizes and carbides at the grain boundaries, which has been found to
be relatively resistant to PWSCC.
Westinghouse Practice. Up until the late 1970s, Westinghouse practice involved use of
relatively low temperature mill-annealed tubing which was not thermally treated [Hunt and
Gorrnan 1986]. For these earlier steam generators, prior to the introduction of improved heat
treatment and other fabrication improvements discussed below, the residual stresses and
microstructure of the tube material are such that the tubes are relatively susceptible to
primary- and secondary-side stress corrosion cracking.
Starting in the late 1970s, Westinghouse used an array of features to reduce the potential
for tube corrosion. These features included thermal treatment of tubing for 15 hours at 705°C
to relieve the residual stresses and improve the microstructure, followed by stress relief of
tight radius U-bends. Improvement of the microstructure involves precipitation of the
chromium carbides at the grain boundaries. In addition, holding the tubing in the precipitation
range for a long period of time allows the chromium to diffuse from the grain interiors to
chromium depleted regions near the grain boundaries, preventing sensitization. Because of the
improvements associated with this thermal treatment, experience with thermally treated Alloy
600 tubing has shown that only a small fraction of it is susceptible to PWSCC in highly
stressed areas.
Current Practice. Current practice by the steam generator suppliers in France, Japan and
the USA is to use thermally treated Alloy 690. This alloy, which is similar to Alloy 600 but
has about twice as much chromium (29.5% rather than 15.5%) and proportionally less nickel,
has been found in tests to be very resistant to primary water stress corrosion cracking and to
have unproved corrosion resistance in secondary-side environments. Most vendors are using
a thermal treatment of about 705°C for 15 hours to relieve the fabrication stresses and
improve the microstructure. Some vendors thermally treat the tight radius U-bends for various
times up to an additional two hours at about 700°C to relieve the residual stresses induced
by bending and peen the inside surfaces of the tube legs to produce a layer of cold worked
material a few tens of microns deep.
Siemens/KWU Practice. The first two Siemens/KWU steam generators were supplied
with Alloy 600 mill-annealed tubing and began leaking after two years of operation.
Thereafter, all Siemens/KWU steam generators were fabricated with Alloy 800M tubing
(33.5% Ni, 21.5% Cr, 44% Fe, < 0.03% C, < 0.6% Ti). Compared to the standard Alloy 800
ASTM specification, Siemens Alloy 800M has a reduced carbon content to minimize
20
sensitization, an increased stabilization ratio (Ti/C > 12), and slightly increased chromium and
nickel contents to achieve a higher resistance to pitting and transgranular stress corrosion
cracking.
CANDUPractice. Following the use of Alloy 600 in a small demonstration reactor, the
material used in the 1960's in the CANDU steam generators was Monel 400 a high
nickel/copper alloy. This alloy has good corrosion properties but is extremely sensitive to
oxygen content. Its ferromagnetic properties also increase the difficulty of inspection with
standard eddy current coils.
The material used for later units was changed to Alloy 600. The practice for Babcock
& Wilcox Canada Ltd. for manufacturing Alloy 600 tubing (high temperature mill-annealing
and heat treatments) was very similar to the practice of its parent company as described for
PWRs. As a result, this type of tubing tends to behave similarly, with respect to degradation
mechanisms, to that used in once through steam generators built by Babcock & Wilcox in the
USA.
The current practice for CANDU RSGs is to use titanium stabilized Alloy 800M tubing
and a manufacturing method which precludes any random heat addition to the tubing.
WWER Tubing Material. The WWER-440 and WWER-1000 steam generator tubing is
made of Type 08X18N10T stainless steel which is a Ti-stabilized austenitic stainless steel
with 0.08% carbon, 18% chrome, 10% nickel and less than 1% titanium.
3.2.2. Tube Installation in the Tubesheet
PWR and CANDU steam generator tubes have been installed in a thick tubesheet and
WWER steam generator tubes have been installed in somewhat thinner walled collectors by
mechanical rolling, hydraulic expansion, or explosive expansion (which may introduce high
residual stresses) and seal welding to the tubesheet/collector inside surface cladding. For the
early PWR plants, the mill-annealed tubing was connected to the tubesheet by hard rolling
the tube into the bottom of the tubesheet for a length of about 60 to 100 mm. This left an
approximately 0.2-mm wide, radial crevice (where chemical impurities could concentrate)
between the tube and tubesheet along the top portion (about 460 mm) of the tubesheet. hi
later steam generators of Westinghouse design (early to mid-1970s), the tubing was expanded
for the rest of the tubesheet height using an explosive expansion process (Wextex expansion)
in the field or by additional hard rolling in the shop, hi cases where the expansion was done
by additional rolling, field experience has shown that high residual stresses were introduced
into some tubes during rolling anomalies, e.g., at regions rolled twice or at transition regions
where rolling was skipped. For Westinghouse type steam generators made in the later part of
the 1970s, full-depth tube expansion was accomplished in the shop using hydraulic methods.
The Siemens/KWU steam generators were fabricated with either a three or two step
mechanical hard roll until the late 1980s. The most recent procedure used by most of the
PWR and CANDU steam generator manufactures is to perform a hydraulic expansion over
nearly the entire tubesheet thickness (stopping and starting within a few mm of each end)
followed by a one (near the top) or two step mechanical hard roll near the top and near the
bottom (called a kiss roll). The transition region is formed by the hydraulic expansion, which
leaves significantly lower residual stresses in the tubing than the hard mechanical roll
21
expansions. The hard mechanical rolling near the top or near both ends of the tubesheet
provides a larger holding force than can be obtained with a hydraulic expansion.
Kiss rolls have been used to install the tubes in the tubesheets of the French steam
generators since 1980. This has resulted in lower residual stresses on the secondary side of
the tubing, but an increased sensitivity to axial cracking on the primary side surfaces.
Westinghouse uses only a hydraulic tube expansion. Westinghouse also machines the
tubesheet faces parallel to within 0.38 mm so that the secondary side crevice depth is less
than 2.5 mm.
Mechanical tubesheet crevices generally do not exist in the CANDU steam generators.
Early units closed the tubesheet crevices by a second roll near the top (secondary side) of the
tubesheet. Current CANDU models use a hydraulic method to close the tubesheet crevice.
The WWER steam generators use two vertical cylindrical collectors or headers, each
with an inside diameter of 800 mm (WWER-440) or 834 mm (WWER-1000) and a wall
thickness of 136 mm (WWER-440) or 171 mm (WWER-1000) rather than a thick-wall
tubesheet. As mentioned above, the WWER-440 collectors are made of the same Ti-stabilized
stainless steel as the tubing. The WWER-1000 collectors are made of the same low alloy
bainitic steel (Type 10GN2MFA) as the vessel, with stainless steel cladding on the inside
surface. The tubes are embedded against the collector wall by explosion or hydraulic
expansion and welded at the collector inside surface using argon-arc welding. Collector-tube
crevices generally do not exist, however, some "under-rolling" of the heat exchanger tubes
into the collector wall has been reported, resulting in crevices with depths up to 20 mm
(explosive expansions) or 2 mm (hydraulic expansions).
3.23. Tube Supports
Several types of tube support systems have been used in PWR steam generators, as
shown in Fig. 7 [EPRI 1985a]. Most of the original steam generators of Westinghouse design
have plate-type tube supports, where tubes pass through drilled holes in the plate. This
construction leaves a narrow gap around the tube, between the tube and plate, which allows
secondary coolant to flow through. Separate smaller holes are also provided for the secondary
coolant flow. Combustion Engineering steam generators mostly use supports formed from a
lattice arrangement of bars (egg-crate tube supports), but also use drilled plates in some
locations in the U-bend region (see Fig. 8). Babcock & Wilcox steam generators have
plate-type tube supports, but the holes are broached to give a non-circular hole with three
lands to support the tube, with a larger diameter between the lands to allow coolant flow
adjacent to the tube (trefoil-design broached hole). Later Westinghouse designs also use
broached hole tube support plates (with four lands to support the tube-quatrefoil design). The
earlier models have carbon steel as the tube support material, whereas the later models have
corrosion resistant Type 405 ferritic stainless steel.
Antivibration bars or plates are used in the U-bend regions of recirculating steam
generator tube bundles to stiffen the tubes and limit vibration amplitudes. Typical
arrangements for antivibration bars in Westinghouse and Combustion Engineering steam
generators are shown in Fig. 8. The antivibration bars in Westinghouse type RSGs are
installed to provide support to at least row 11, though many were installed to deeper depths,
22
Vertical
hot/cold leg
supports
VH
Outermost
tube
cio
Diagonal bar Diagonal bar
09
hotteg coldteg
C8
Innermost
tube
Hot leg
supports H4
Tubesheet
Superheater
section
CE recirculating steam generator
Tubesheet
Pnmsuy
outlet
nozzles
HTS CTS B&Wonce-through

steam generator (3)
Intel Outlet
Westinghouse recirculating
steam generator
Tube Support Plate/Tubesheet Nomenclature
WMinghous* model 51
FIG. 7. Typical steam generator tube support layouts -with tube support plate and tubesheet
nomenclature.
23
Vertical tube support bar
(bat wing)
Row 73
No. 8 tube support
partial drill hole
support plate
No. 7 tube support

partial egg crate
with drill hole
plate section
No. 6 tube support

full egg crate
CE RSG Anti-vibration with drill hole
bar arrangement plate section
AVB Tube
7/8 in.
Model F Anti-vibration
11 16 bar arrangement
Model 51 Anti-vibration
bar arrangement
FFG. 8. Typicd recirculating steam generator antivibration bar arrangement.
e.g., to row 8. The antivibration bars in later Westinghouse models have a square cross
section and are made from Alloy 600 and are chrome plated. The arrangement of antivibration
bars in Combustion Engineering steam generators includes vertical, horizontal, and bat wing
strips, as shown in Fig. 8.
The CANDU steam generator tube support design has gone through many changes. Older
operating units have a carbon steel lattice grid arrangement, or carbon steel trefoil broached
plates (see description of Babcock & Wilcox design above). Recent models use an advanced
version of the lattice grids made of stainless steel (see Fig. 9). Antivibration (U-bend)
supports have also undergone changes, from carbon steel scallop bars (stacked and staggered)
to the current stainless steel flat bar type.
The WWER-440 and WWER-1000 steam generators use stainless steel bar and stamped
wave-like plates to separate and support the tubes.
24
Tube-, Plate-,
L
— Tube to Support
Clearance
High Support Low Support

r Bar r Bar
I— Support Bar
Row Space—I
Free Row Space Between B&W Lattice Grid

and Broached Rate Design
FIG. 9. Typical CANDU steam generator tube support structures. (Courtesy of C. Maruska,
Ontario Hydro.)
25
/=\
Feedwater
nozzle
Stub barrel weld
MS 10456
FIG. 10. PWR steam generator showing shell welds (Westinghouse 1990). Copyright
Westinghouse Electric; reprinted with permission.
Nozzle Mend
radius cracks
FIG. 11. Feedwater nozzle sites susceptible to high-cycle thermal fatigue damage caused by
turbulent mixing of leakingfeedwater and hot steam generator coolant (Westinghouse 1990).
Copyright Westinghouse Electric; reprinted with permission.
26
3.2.4. Feedwater Nozzle and Shell
Figure 10 shows the locations of the feedwater nozzle and the girth welds in a schematic
of the shell of a Westinghouse PWRrecirculating steam generator without a preheater. Figure
11 shows a typical Westinghouse feedwater nozzle and thermal sleeve. The Westinghouse
thermal sleeve is welded to the feedring (not shown in figure). It fits snugly against the
nozzle, but is not attached to the nozzle. Figure 12a shows the original configuration for the
piping-to-nozzle weld (see Section 8.6 for repair). The steam generator shell, including the
feedwater nozzle, is made of low-alloy ferritic steel, typically SA-533 Type A, Class 1 or 2
for the Westinghouse steam generator shells and SA-508 C12 for the feedwater nozzle
forgings. (Some of the earlier steam generators made by Westinghouse in their Lester plant
used SA-302 Grade B for the plate material, but all the steam generators built at the Tampa
plant used SA-533.) The thermal sleeve inside the feedwater nozzle is made of SA-106
Grade B carbon steel.
As stated in Section 3.2.1, Babcock & Wilcox heat treated the entire Babcock & Wilcox
once-through steam generator at about 595°C for 15 hours, thus reducing residual stresses in
the shell and feedwater nozzle, as well as in the tubing. Most of the other steam generator
vendors did not heat treat the entire steam generator.
The WWER steam generator pressure vessels and feedwater nozzles are shown in Figs
4, 5, and 6. The WWER-440 steam generator shell is made of Type 22K carbon steel. The
WWER-1000 steam generator shell and feedwater nozzle is made of Type 10GN2MFA low
alloy steel with the following chemical composition: 0.08% to 0.12% carbon, 0.17% to
0.37% silicon, 0.8% to 1.1% manganese, < 0.30% chromium, 1.8% to 2.3% nickel, 0.4% to
0.7% molybdenum, 0.03% to 0.07% vanadium, and less than 0.02% sulphur and phosphorus.
Weld
! 1 17mm
1
28 mm (1.1 in.) (0.671 in.)
Deepest^ < \
crack ^Vl
v ti/
\
\
/
/
/
17mm
. (0.656 in.)
| k I ^**«--_ _ . \
I I ^^
Backing strip \
Nozzle end
[406 mm (16 in.) Schedule 80] I406mm <16in'>Schedule "»
(a) Original design
(b) Typical repair configuration V
Pipe end / Nozzle end

[406 mm (16 in.) Schedule 80] [406 mm (16 in.) Schedule 60]
FIG. 12. Crack locations in the D.C. Cook nozzle (USNRC 1980).
27
4. STEAM GENERATOR DEGRADATION MECHANISMS
This section discusses the stressors, susceptible sites and failure modes associated with
the various steam generator degradation mechanisms. PWR and CANDU recirculating steam
generator tube degradation is discussed first, including primary water stress corrosion
cracking, outside diameter stress corrosion cracking, fretting, pitting, denting, high-cycle
fatigue, and wastage. This material is followed by similar information on PWR once-through
steam generator tube and WWER steam generator tube degradation. A discussion of the tube
rupture experience is then presented, followed by information on PWR steam generator shell
and feedwater nozzle degradation and WWER collector stress corrosion cracking and
feedwater system erosion-corrosion. Steam generator plug and sleeve degradation is discussed
in Section 8.
4.1. PWR AND CANDU RECIRCULATING STEAM GENERATOR TUBES
The relative impact of tube degradation mechanisms on overall PWR steam generator
performance has dramatically changed over time. Figure 13 shows the percentage of the total
number of tube failures1 caused by each of the major degradation mechanisms for the years
1973 through 1994 (EPRI 1995a). Both PWR and CANDU recirculating steam generator and
PWR once-through steam generator tube failures world wide are included (Figure 13 does
not include data from the WWER reactors, except Loviisa Units 1 and 2.) Phosphate wastage
was the major cause of tube failures in PWR steam generators until about 1976. From 1976
to about 1979, denting was the major cause of PWR steam generator tube failures. After
about 1979, a variety of corrosion mechanisms became important, including intergranular
stress corrosion cracking/intergranular attack and pitting on the outside diameters of the tubes
and primary water stress corrosion cracking on the inside surfaces. Fretting damage became
more apparent after about 1983.
Table V lists the number of PWR and CANDU plants reporting various problems in
1977, 1982, and 1993 (EPRI 1994). There was a dramatic increase over the last 15 years in
the number of plants reporting primary water stress corrosion cracking, outside diameter stress
corrosion cracking, and fretting problems. Over 50% of the PWR units world wide have now
reported some occurrence of tube fretting and wear. However, some plants report no
problems, even after five years of operation (7-10% of the plants report no problems after five
years of operation). Most of the PWR steam generator tubes which have failed over the years
have been mill-annealed Alloy 600 tubes. However, some failures of thermally treated Alloy
600 tubing have been reported, primarily due to fretting and denting (degradation mechanisms
due to the design of the support plates and antivibration bars and the presence of loose parts,
rather than the tubing material). But there have also been a few failures of thermally treated
Alloy 600 tubing due to primary and secondary side stress corrosion cracking.
1
Failure is defined as a nondestructive examination (NDE) indication requiring the tube to be removed from
service (plugged) or repaired. The tubes that actually leaked primary coolant are a small proportion of the tubes
plugged or repaired. Steam generator tubes are sometimes plugged as a preventive action if they are judged to
have a high probability of future failure.
28
100
73 74 75 76 77 78 79 80 B1 82 83 84 85 86 87 88 89 90 91 92 93
FIG. 13. Worldwide causes of steam generator plugging (EPRI 1995a). Copyright 1996
Electric Power Research Institute; reprinted with permission.
The Alloy 800M tubing used in the Siemens/KWU steam generators has performed
relatively well. There were Alloy 800M tubing failures due to wastage in the Siemens/KWU
steam generators which began operation in the 1970s with phosphate water chemistry, but
there have been no wastage failures in the Siemens/KWU steam generators which began
operation in the 1980s with an all volatile water treatment. There have also been some Alloy
800M tubing fretting failures in the Siemens/KWU steam generator which began operation
before 1986. But, only one Alloy 800M pulled tube has exhibited a stress corrosion crack, pits
have been found on only two Alloy 800M tubes, and no Alloy 800M tubes have exhibited
detectable intergranular attack or primary water stress corrosion cracking. There have been
no Alloy 690 tube defects of any kind reported to date.
Figures 14 and 15 identify degradation sites for PWR and CANDU steam generators,
respectively. Table VI lists PWR steam generator degradation mechanisms, sites, stressors,
failure mode and inspection methods for tubes and tubesheets. Table VII lists the degradation
mechanisms and sites currently active in the CANDU steam generators and the corresponding
countermeasures completed or in progress.
4.1.1. PtinBiy Water Stress Cbnosion Cracking (PWSCQ
Research has shown that stress corrosion cracking of austenitic stainless steels and nickel
based alloys requires at least the following three conditions: susceptible tubing microstructure
(alloy content or few intergranular carbides), high applied or residual tensile stress (near the
yield strength), and a corrosive environment (high temperature water). The influence of the
nickel content on the stress corrosion cracking processes in 18% chromium austenitic alloys
when stressed slightly above the yield point of the material in demineralized water or water
containing 1 g/L chloride ions is shown in Fig. 16 (Berge 1993). As indicated on the figure,
Alloy 600 can be susceptible to pure (primary) water stress corrosion cracking, where as,
29
TABLE V. UNITS REPORTING STEAM GENERATOR PROBLEMS (EPRI 1994)
DATE 3/77 8/82 12/93

NO. UNITS: 52 99 235
REPORTED PROBLEMS:
Denting
- Tube Support Corrosion 15 30 36
- At and Above the Tubesheet 6 12 50
Tubing Corrosion
- Wastage 19 28 39
- Pitting 0 3 16
- PWSCC 1 22 102
- ODSCC/IGA 6 22 85
Mechanical Damage
- Fretting 9 15 128
- Fatigue Cracking 3 4 15
- Impingement 0 2 10
No Problems 26 32 48
No problems after 5 years ops (no. of units/no. >5yrs. ops). 1/14 4/57 20/205
Units reporting no problems after five years of operation
3/77 8/82 12/93
Trillo Kewaunee Bruce-B5 Obrigheim (Rpl)
Mihama3 Brokdorf Philippsburg 2
Neckarwestheim ChinonB 3 Pickering-A 3
Davis Besse CruasB Pickering-A 4
Cruas4 Pickenng-B 7
Genkai2 Pickering-B 8
Grafenrheinfeld Robinson 2 (Rpl)
Isar2 Trillo 1
Loviisa 1 Uljinl
Loviisa 2 Wolsung 1
30
Inner row Fretting or
U-bends AVBwear
High-cycle
fatigue at
dents
All tube
support locations
on hot leg side
IGA/IGSCC
onOD,
Sludge
Hot and cold leg

below 1st support
IGA/IGSCC
on 00
PWSCC
cracking
on ID
Hot&coUteg Hot leg

tube-to-tubesheet crevices tube-to-tubesheet crevice
Top of ret expansion
FIG. 14. Locations of known tube wall degradations in recirculating steam generators.
(Courtesy ofK. J. Krzywosz of the Electric Power Research Institute NDE Center; modified.)
Alloy 690 and Alloy 800M are generally not susceptible to PWSCC. Austenitic stainless
steels with a nickel content below about 15% are susceptible to transgranular stress corrosion
cracking when exposed to water containing significant amounts of chlorides (1 g/L). The
occurrence of PWSCC of Alloy 600 strongly depends on the absence of intergranular
carbides. High mill-annealing temperatures (1065°Q during final heat treatment produce
intergranular carbides, which make Alloy 600 tubes resistant to PWSCC. In contrast, low
mill-annealing temperatures produce intragranular carbides, which make tubes susceptible to
PWSCC. PWSCC is a stress-dependent process such that the damage rate increases as the
stress to the fourth power. The exponent of four on damage rate is typical of stress exponents
for creep and, thus, is consistent with modern models for PWSCC, which say that slow
straining at the crack tip is an essential part of the cracking process. PWSCC is also a
thermally activated process, which can be described by an Arrhenius relationship. A small
decrease in steam generator operating temperature will significantly slow the initiation and
growth of PWSCC at any location in the steam generator.
31
Steam Carry Over
Tube High Cycle Fatigue
Corrosion of Supports
Tube Fretting
Tube Corrosion at Supports

- ODSCC (Including Lead Assisted)
Bow Blockage of Supports • Tube Deformation
Secondary Flow Instabilities
Fouling of Tube Surface (10 and OD)

Loss of Heat Transfer Efficiency
Tube Pitting in Freespan Area

Tube Denting at Supports
— Tube Corrosion in Sludge Zone Above

Tubesheet and Tube Supports
Loose Parts Damage — •Pining
-IGA
Primary Side Plate Leakage
FIG. 15. Degradation mechanisms and locations in CANDU recirculating steam generators.
(Courtesy of C. Maruska, Ontario Hydro.)
PWSCC occurs at locations on the inside surfaces of recirculating steam generator tubing
with high residual stresses (introduced during fabrication and installation of the tubes, as
discussed in Section 3.2). These locations are primarily the roll-transition regions in the
tubesheets, the U-bend regions of the tubing in the inner rows (i.e., the tubes with a small
bend radius), and any dent locations at the tube support plate, tubesheet, or sludge pile
elevations. Section 4.1.5 discusses tube denting, e.g., deformation resulting in residual stresses
due to buildup of corrosion products. PWSCC generally occurs on the hot-leg side of the
recirculating steam generations; however, cold leg PWSCC has been observed.
32
I ABLE VI SUMMARY OF PWR RECIRCULATING STEAM GENERATOR TUBE DEGRADATION PROCESSES
Rank3 Degradation Stressor Degradation Sites Potential Failure ISI Method

Mechanism Mode
1 ODSCC Tensile stresses, impurity • Tube-to-rubesheet crevices Axial or circumferential crack MRPC
concentrations, sensitive • Sludge pile Circumferential crack MRPC/Cecco5
materials • Tube support late Axial crack Bobbin coil/Cecco 5
• Free span Axial crack Bobbin coil (in absolute mode)
2 PWSCC Temperature, residual tensile • Inside surface of U-bend vlixed Crack MRPC*
stresses, sensitive materials (low • Roll transition w/o kiss rolling viixed Crack MRPC
mill anneal temperature) • Roll transition with kiss rolling Axial Crack MRPC
• Dented rube regions Circumferential Crack Bobbin coil or MRPC
3 Fretting, ;low induced vibration, • Contact points between tubes and the AVBs, or tubes and the jocal wear Bobbin coil
Wear aggressive chemicals preheater baffles
• Contact between tubes and loose parts Depends on loose part geometry Bobbin coil
• Tube-to-tube contact Axial Wear Bobbin coil
4 High-cycle Iigh mean stress level and flow At the upper support late if the rube is clamped Transgranular circumferential Leak detection or by detection
fatigue induced vibration, initiating cracking of precursor
defect (crack, dent, pit, etc )
5 Denting Oxygen, copper oxide, At the rube support plates, in the sludge pile, in the tubesheet Flow blockage in tube, may lead to Profilometry, bobbin coil
chlorides, temperature, pH, crevices circumferential cracking (see
crevice condition, deposits PWSCC), decreases the fatigue
resistance
6 Pitting Brackish water, chlorides, Cold leg in sludge pile or where scale containing copper deposits is Local attack and tube thinning, may Bobbin coil, ultrasonics
sulphates, oxygen, copper found, under deposit pitting in hot leg lead to a hole
oxides
7 Wastage 'hosphate chemistry, chloride rubesheet crevices, sludge pile, rube support plates, AVBs General thinning [Bobbin coil
concentration, resin leakage
"Based on operating experience and number of defects (as of 1993)

b
Multifiequency rotating pancake coil probe
TABLE VH. SUMMARY OF MAJOR CANDU STEAM GENERATOR DEGRADATION
Rank Degradation Stressor Degradation Sites Potential Failure ISI Method

Mechanism Mode
1 ODSCC Lgh stress, corrosive J-bend support Circumferential Cecco3
environment due to intersections cracking
ieposit build-up
J
7th support plate redommantly Cecco 3
circumferential, some
axial
2 Outside diameter Deposits which cause Tubesheet area under jocal tube thinning E/C [carter]
pitting a corrosive sludge pile and at eading to holes Ultrasonics
environment lower tube support
intersections and at
freespan tubes
3 Fretting Flow induced Q-bend support detal loss which Bobbin coil
vibration, loose intersections may lead to large
supports hole
4 Corrosion of carbon Corrosive U-bend supports Support Visual (secondary
steel supports environment, stress disintegration and access)
metal loss/may lead
to tube degradation
from flow induced
vibration due to lack
of support
5 Erosion-corrosion, Bolt failure Primary head freak up of bolted Visual and
high or low cycle dates may lead to metallography of
fatigue >lockageofPHT bolts
inlet
Susceptibility to SCO
Transgranular Intergranular
Stress Corrosion Cracking (SCC)
Cell-containing Water Demineralized Water

(1000ppm)
1 1 1 1 1 I 1
L
3 10 20 30 ' 40 50 1 60 70' Ni-Content
Inconel 800 Inconel 690 Inconel 600
19...23% Cr 27...31%Cr 14...17%Cr
FIG. 16. Schematic diagram showing the influence of nickel content on the cracking processes
occurring in three steam generator tubing materials stressed slightly above the yield point in
350°C water. (From Coriou 1971, as reported by Berge 1993. Copyright 1993 Electric Power
Research Institute; reprinted with permission.)
34
Examination of removed tubes affected by PWSCC and in situ inspection by rotating
pancake coil eddy-current test probes indicate that PWSCC cracks typically have the
following patterns (Dobbeni et al. 1985, Engstrom 1985):
1. Cracks in U-bends typically are axial in orientation, though occasional off-axial cracks
have been detected.
2. Cracks in standard roll transitions are mostly axial, though occasional short
circumferential cracks occur between axial cracks. Rarely are isolated circumferential
cracks detected. When the tube expansion is kiss rolled the crack is always axial and the
growth rate decreases as crack length increases.
3. Some large circumferential cracks have been detected in the sludge pile area of
kiss-rolled plants in France. In some cases, large circumferential cracks have been
located at the same transition as multiple axial cracks. In other cases, large
circumferential cracks have occurred without any axial cracks.
4. Cracks at explosive transitions are typically circumferential in orientation, though
occasionally axial PWSCC is noted by rotating pancake coil eddy-current testing.
5. Primary-side cracks at dented tube support plate intersections are typically axial, though
some circumferential segments have been noted.
6. Cracks at dents associated with sludge pile deposits at the top of the tubesheet (observed
in France) have been circumferential in orientation.
In the case of an axial crack, a leak will occur before the critical crack size (leading to
tube rupture) is achieved. On the other hand, the evolution of circumferential cracks is not
known. Consequently, a tube with a circumferential crack is usually plugged or sleeved
immediately after detection to avoid possible tube rupture.
As of December 1993, at least 94 PWR plants worldwide (36 plants in the USA) with
RSGs had experienced significant PWSCC at the roll transition (tubesheet), dent, and/or U-
bend locations of the tubing (EPRI 1994). Approximately 14 180 recirculating steam generator
tubes with PWSCC at or near the roll transitions have been plugged at 85 plants. Tubes with
PWSCC have also been sleeved at 17 plants. Approximately 8430 recirculating steam
generator tubes with PWSCC in the U-bend regions have also been plugged at 63 plants
(however, several hundred tubes were preventatively plugged and may not have been
defective). Fifty-three PWR plants have experienced both transition region and U-bend
PWSCC and tubes with PWSCC at dents have been plugged at, at least, five plants.
This degradation has occurred primarily at Westinghouse-type plants (steam generators

built by Westinghouse and by Westinghouse licensees in Europe and Japan) in Belgium,
France, Japan, the Republic of Korea, Spain, Sweden, Switzerland, and the USA with Alloy
600 mill-annealed tubing, typically expanded by hard rolling (which introduces high residual
stresses at the roll transition and where rolling anomalies occurred). The most extensively
degraded steam generators have had as many as 20% to 38% of all their tubes plugged as a
result of PWSCC and have been replaced at some plants. However, similar steam generators
(same model number) at other PWR plants have experienced only a few tube failures due to
PWSCC.
35
The Combustion Engineering plants with relatively high-temperature mill-annealed tubing
initially reported significantly less PWSCC. However, both explosive-transition and U-bend
PWSCC occurred at Maine Yankee after about 16 years of operation (model CE-67 steam
generators) and a few cracks at the roll transition regions occurred at Palo Verde Units 1 and
3 after five to seven years of operation (model CE-80 steam generators). Recently, it was
reported that after 22 years of operation 60% of the Maine Yankee steam generator tubes had
indications of circumferential cracking at or near the expansion transition and the utility
sleeved all 17 000 tubes in its three steam generators (USNRC 1995).
Nine French plants with thermally treated Alloy 600 tubing have plugged tubes because
of PWSCC at the roll transition region. (PWSCC of thermally treated Alloy 600 tubing has
occurred only in steam generators in which the tubes were mechanically rolled into the tube
sheet.) However, the number of steam generator tubes involved (a total of 82) is rather small,
in part, because the tubes were not plugged unless they also had dents (i.e. the possibility of
circumferential cracking). As of December 1993, there had been no PWSCC of thermally
treated Alloy 600 tubes in the USA or elsewhere outside of France and there had been no
PWSCC of Alloy 690 or Alloy 800M tubing.
CANDU units using high-temperature mill-annealed Alloy 600 tubing, running at

relatively low inlet temperatures, have not experienced PWSCC to date. The oldest running
plant with Alloy 600 tubing has over 11 effective full power years (EFPY) or approximately
18 calendar years of operation with no evidence of this degradation mechanism. It is believed
that the lower operating temperatures of the CANDU primary system and more resistant
material may have contributed to the delay in onset of this type of cracking. The other tubing
alloys used in the CANDU steam generators, Monel 400 and Alloy 800, are not susceptible
to PWSCC.
4.1.2. Outside Diameter Stress Corrosion Cracking (ODSCQ
Outside diameter stress corrosion cracking is a degradation mechanism which includes

both intergranular stress corrosion cracking (IGSCQ and intergranular attack (IGA) on the
outside surfaces of the tubing. Most of this degradation takes place in the tube to tubesheet
and tube to tube support plate crevices, however, sludge pile and free-span ODSCC has been
observed at some plants. IGSCC requires the same three conditions as PWSCC: tensile stress,
material susceptibility, and a corrosive environment (in this case, high temperature water
containing aggressive chemicals). IGSCC cracks occur along the grain boundaries, oriented
normal to the maximum principal stress. IGA is characterized by local, corrosive loss of
material on the grain boundaries; it does not require large tensile stress, but it is believed that
stress accelerates initiation and growth of this mechanism
ODSCC strongly depends on the concentration of corrosive impurities at dryout regions

in the steam generator. The impurity levels in secondary-side systems are highly variable, and
are likely influenced by at least the following: crevice geometry, cooling water type (fresh,
brackish, sea), secondary plant materials (e.g., presence of copper), condenser leakage history,
air in-leakage history, water treatment history, plant attention to secondary-side chemistry, and
types and application history of remedial measures. ODSCC has occurred in the tube to
tubesheet and the tube to tube support plate crevices, in the sludge pile region, and at free-
span locations, especially free-span locations with significant crud build-up. Most outside
36
diameter stress corrosion cracks are primarily oriented in the axial direction, however,
significant circumferential cracking has been observed in the roll transition region of the
tubing in some steam generators and circumferential ODSCC is sometimes found near dents.
For example, circumferential ODSCC at the roll transition region of the tubing had occurred
in about 50% of the tubes in each of the three steam generators at Doel Unit 4 by the end of
the 8th cycle. The axial cracks occur either singularly or in networks of multiple cracks,
sometimes with limited patches of IGA Shallow circumferential cracks may sometimes occur
in the IGA affected regions producing a grid-like pattern of axial and circumferential cracks
termed "cellur corrosion."
As of December 1993 at least 89 PWR plants (44 US plants) with RSGs and a few
CANDU plants have experienced some degree of ODSCC in the tubesheet crevice, sludge
pile, tube support plate intersection, or free-span locations (EPRI 1994). Approximately 14
140 recirculating steam generator tubes with ODSCC at the tube support plate locations have
been plugged at 63 PWR plants. Approximately 13 860 recirculating steam generator tubes
with ODSCC in the tubesheet crevice and sludge pile regions have also been plugged at 75
PWR plants (49 PWR plants have had both tube support plate and tubesheet ODSCC repairs).
Tubes with ODSCC have also been sleeved at 25 plants. This degradation has occurred
primarily in Combustion Engineering (eight plants) and Westinghouse-type plants (79 plants)
with Alloy 600 mill-annealed tubing. Only one tube with ODSCC has been found in the
Siemens/KWU steam generators with Alloy 800M tubing and only one plant with thermally
treated Alloy 600 tubing has reported ODSCC (Kori-2 has reported finding ODSCC in the
tubesheet region and plugging 125 tubes). The only CANDU plant with extensive ODSCC
has been Bruce-A2 where 1399 tubes failed (were plugged) due to lead assisted stress
corrosion cracking. The most extensively degraded steam generators have had as many as 40
to 56% of all their tubes plugged or sleeved as a result of ODSCC and have been replaced
at a number of plants (or in some cases, the plants have been shut down). However, similar
steam generators (same model number) at other plants have experienced only a few per cent
failures due to ODSCC.
ODSCC has appeared in PWR steam generator tubes with both high and low
mill-annealed temperature, but generally not in thermally treated tubes because the thermally
treated tubes do not have chrome depletion at the grain boundaries. Tests were conducted
using high-temperature electrochemical measurements to identify conditions leading to IGA
[Pinard-Legry and Plante 1983]. The results of these tests indicate that in 10% caustic media
at 320°C, IGA is commonly observed in Alloy 600 in the mill-annealed condition.
Thermally-treated material at 700°C shows definite improvement over mill-annealed material
in resistance to both IGA and IGSCC.
The oldest running CANDU units tubed with Alloy 600 are currently experiencing
widespread, but relatively shallow (5-10% of the wall thickness) ODSCC at the tube U-bend
supports. This degradation has been due to a combination of heavy secondary side deposits
which created an aggressive environment on the tube surface and corrosion of the carbon steel
supports which caused high stresses in the area. The heavy deposits in the steam generators
were due to early water treatment plant problems, condenser leakage, abnormal chemistry
incidents, and feedtrain corrosion problems. Final failure of a few tubes occurred due to high
cycle fatigue.
37
The degradation was severely aggravated in Bruce-A2 by contamination due to a lead
blanket inadvertently left in one steam generator during maintenance activities. The lead was
transported into the other steam generators at the unit through the water in the common steam
drum. Cracking in the lead contaminated steam generators was typical of lead assisted
cracking: mixed mode, transgranular and intergranular, ranging from 0-100% throughwall.
Lead shielding was also inadvertently left behind in the Doel Unit 4 Steam Generator B in
Belgium and is believed to have contributed to the severe ODSCC which subsequently
occurred in that steam generator.
Because ODSCC can take several forms (short axial cracks, long axial cracks,
circumferential cracking, cellur corrosion, etc.) and the ease of detection of these various
kinds of ODSCC degradation varies considerably, the potential safety consequences of
ODSCC at separate plants can be quite different. For example, ODSCC within the tubesheet
is much more difficult to detect with a standard eddy-current bobbin coil probe than PWSCC
within the tubesheet or axial ODSCC at the tube support plates. However, it is possible to
detect ODSCC within the tubesheet before it reaches a critical size and, therefore, make
repairs before tube rupture. To date, there have been no tube ruptures due to undetected
ODSCC in the tubesheet region. Axial ODSCC at the tube support plates can usually be
readily detected with a bobbin coil probe, however, detection of circumferential ODSCC at
the tube support plates requires special probes as does the sizing of ODSCC. Also, the
evolution of the ODSCC depends significantly on the local environment within the crevice
or under the crud, the details of which are often unknown. Therefore, the future crack growth
rate cannot always be accurately estimated. However, some tube supports (and the tubesheet)
can provide reinforcement in the event of a throughwall crack, provided the support does not
move relative to the tube during the event and the crack is within the support. Freespan
IGA/IGSCC can occur if there are deposits on the tube, which concentrate impurities. The
sensitivity of the eddy-current signal is poor and a special analysis in absolute mode is needed
to detect a freespan flaw before the flaw achieves a critical size. Tube ruptures have occurred
due to freespan ODSCC.
4.13. Fretting, Wear and Thinning
These steam generator degradation types are broadly characterized as mechanically-

induced or -aided degradation mechanisms. Degradation from small amplitude, oscillatory
motion, between continuously rubbing surfaces, is generally termed fretting. Tube vibration
of relatively large amplitude, resulting in intermittent sliding contact between tube and
support, is termed sliding wear, or wear. Thinning generally results from concurrent effects
of vibration and corrosion. However, thinning occurs at some locations where flow-induced
vibrations are not expected, so it is not certain that tube motion is required for this
mechanism; in some cases it may simply be the result of corrosion wastage. Fretting and wear
makes tubes susceptible to fatigue crack initiation at stresses well below the fatigue limit,
resulting in through-cracks or tube rupture.
The major stressor in fretting and wear is flow induced vibration. Initiation, stability, and
growth characteristics of damage by these mechanisms may be functions of a large number
of variables, including the support locations, the stiffness of the supports, the gap size
between tube and support, secondary flow velocities and directions, and oxide layer
characteristics.
38
Fretting/wear/thinning degradation was first identified as a problem in about 1973 and
has been noted to some degree in all major PWR steam generator designs. This includes
preheater and antivibration bar (AVB) wear/fretting in Westinghouse-type RSGs, cold leg
thinning in Westinghouse-type RSGs, antivibration bar (diagonal support) wear/fretting in
Combustion Engineering RSGs, and AVB wear/fretting in Siemens/KWU steam generators
(EPRI 1985a).
As of December 1993,116 plants with RSGs had experienced tubing failure due to AVB
wear/fretting, 78 plants had reported wear/fretting failures due to loose parts damage, and 12
plants had reported wear/fretting failures associated with the steam generator preheaters (EPRI
1994). 4633 tubes have been plugged because of AVB wear/fretting (920 tubes were
preventatively plugged and the rest were plugged due to NDE indications), mostly in
Westinghouse-type steam generators. This damage has occurred in the more recent
Westinghouse steam generator designs at 17 plants (Westinghouse models F, 44F, and 5IF
and Mitsubishi Heavy Industries model 5IF) as well as in the earlier model steam generators.
Steam generator tubes have also been plugged due to AVB (batwing or vertical strap)
wear/fretting at, at least seven Combustion Engineering designed plants, three Siemens/KWU
plants and one CANDU plant. At least 941 tubes have been plugged because of loose parts
damage in 78 plants, although most of these plants (44 plants) have plugged less than ten
tubes each. One plant did plug 176 tubes due to loose parts damage.
Fretting is occurring in CANDU steam generators with U-bend supports made of

staggered scalloped bars (the U-bend support bar stack is split into two offset stacks as shown
in Fig. 17, and the tubes are held in semicircular holes). This degradation is caused by flow
i—Tubes
Scalloped Bare
FIG. 17. CANDU'recirculating steam generators staggeredscdlop bar arrangement. (Courtesy

of C. Maruska, Ontario Hydro.)
39
induced vibration of the tubes which is due to U-bend supports which were widely spaced and
perhaps insufficiently rigid. Although tube fretting is severe and widespread, no tube failures
have occurred in CANDU steam generators due to this degradation mechanism to date. There
is evidence to indicate that the fretting rate in these steam generators is decreasing with time,
suggesting that this mechanism is self limiting.
A special case of this type of damage is loose parts damage. Loose parts and other
debris have been found on the secondary side of the steam generators at a large number of
PWRs over the years. These parts include tools (for example, a 152 mm flat file at Wolf
Creek, a grinder wheel at Watts Bar Unit 1, a weld rod at Turkey Point 4, parts of a pocket
knife at D.C. Cook Unit 1, and a 152 mm C-clamp at Point Beach), valve and pump parts (for
example, a check valve pin at Turkey Point Unit 4), equipment used for previous inspections,
broken steam generator material, debris left from previous modifications and repairs (for
example, pieces of steel plate, copper tubing, weld material, wire, etc.), and other things.
These loose parts have been implicated in at least two tube rupture events in operating plants
in the USA. In addition to tube ruptures, loose parts in the secondary side of a number of
plants have resulted in tube damage, and plugging.
One of the worst examples of this problem occurred at Ginna from 1975 to 1982.
Foreign objects including various size pieces of carbon steel plate up to about 150 mm in
length fell onto the tubesheet outside the periphery of the tube bundle during steam generator
modifications performed in 1975 and later. This debris then impacted on the exposed
peripheral tubes during subsequent operation and caused defects. The damaged tubes were
plugged as a result of eddy current indications and/or small leaks. However, the debris
continued to damage the plugged tubes and eventually caused the tubes to collapse and in
some cases to become completely severed near the top of the tubesheet. The severed tubes
and debris then interacted with the adjacent inboard tubes, resulting in fretting type wear of
the adjacent tubes. These tubes, in turn, were plugged as a result of eddy current indications
or leaks. However, damage continued until they also became severed. Eventually, an
unplugged tube in the third row in from the outside row was subjected to fretting type wear
over about 150 mm of length and burst. The wear removed about 84% of the wall thickness
over about 100 mm of length, which caused a relatively long "fish mouth" type burst. The
peripheral tube damage mechanisms were primarily mechanical and included impact, collapse,
fatigue, fretting type wear, abrasion, and ductile overload and tearing.
Corrective measures included removing the debris and stabilizing (staking) and plugging
nearby damaged tubes. USNRC Generic Letter 85-02 requested the US PWR owners to
perform visual inspections in the vicinity of the tubesheet along the entire periphery of the
tube bundle and the tube lane to identify and remove any foreign objects. Such an
examination should be done after any secondary side repairs. Obviously, all tools and
equipment going into a steam generator during an inspection should come out.
Although most loose parts damage has occurred on the secondary side of the steam
generators, there have also been cases of primary side damage, mainly to protruding tube ends
and tube-to-tubesheet welds.
40
4.1.4. Pitting
Pitting is a steam generator degradation type appearing as groups of small-diameter wall

penetrations resulting from local corrosion cells, probably promoted by the presence of
chloride or sulphate acids. Condenser leaks and leakage of beads, resin fines, or regeneration
chemicals from ion exchangers can introduce impurities such as chlorides and sulphates,
which result in local acidic conditions conducive to pitting. Oxidizing conditions and the
presence of copper are probable accelerators. Any barriers to diffusion such as sludge
accumulation on the tube wall will accelerate the pitting process by enhancing chemical
concentration.
Pitting generally occurs at the top or within the cold leg sludge pile region. Pitting
corrosion typically occurs in locally weak spots in the passivated surface of the Alloy 600
tube. These susceptible locations may be the result of localized cold work of the metal, the
presence of metal carbides, sulphides, or other secondary phase particles, or emergence of
grain boundaries at the metal surface.
Significant pitting was first reported in an operating PWR steam generator about 1981.
As of December 1993, only 11 PWR. plants with RSGs had plugged tubes because of pitting
and a few other plants had reported minor pitting degradation of 15% throughwall depth or
less (EPRI 1994). However, a few plants have experienced significant pitting degradation
including Indian Point Unit 3 (1290 tubes plugged because of pitting in the original steam
generators and 3606 tubes sleeved), Kori Unit 1 (804 tubes plugged because of pitting and
1578 tubes sleeved for various reasons), and Millstone Unit 2 (1655 tubes plugged because
of pitting in the original steam generators and 5164 tubes sleeved for various reasons). Most
of this degradation occurred in the cold leg and cold leg sludge pile regions, however, pitting
has also been found on the hot leg side of the RSGs (Laskowski and Hudson 1986, Angwin
1984, Theus and Daniel 1984). In addition, most of the pitting has been associated with Alloy
600 tubing exposed to severe secondary-side chemistry incursions. However, 332 Type 304
stainless steel steam generator tubes at the Yankee Rowe plant were also plugged because of
pitting degradation.
Severe pitting has been experienced in CANDU units tubed with Monel 400 (1994 tubes
plugged at one unit). This pitting is more accurately described as tube outside surface
underdeposit corrosion and is caused by heavy secondary side deposits, both on top of the
tubesheet and in the lower tube support areas. These deposits concentrate aggressive species
such as chlorides and sulphides present due to condenser leakage and water treatment
problems. One CANDU unit tubed with Alloy 800M alloy has experienced a small number
of tube failures due to pitting (underdeposit chloride pitting) at the first and second support
plates. Early condenser tube leakage (seawater) and sludge deposits contributed to this
degradation.
4.1.5. Denting
Denting describes the mechanical deformation or constriction of the tube at a carbon

steel tube support plate intersection caused by the buildup of deposits and the growth of a
voluminous support-plate corrosion product in the annulus between the tube and support plate.
41
Dents do not themselves result in tube wall penetration or reduction in wall integrity.
However, denting at some plants in the past has been sufficiently severe that it caused
structural damage to the tube supports. Denting is a concern because even small dents can
induce tensile stresses above yield strength in the tube wall. As a result, these tubes may be
subject to PWSCC or IGSCC at the dents during subsequent operation. In addition, severe
denting in tubes with tight radius U-bends has accelerated stress corrosion cracking in the
U-bends from distortion of the tube legs [EPRI 1985a, Clark and Lewis 1985]. Also, tubes
with dents at the top tube support plate in the U-bend region of the RSGs are more
susceptible to high-cycle fatigue failure.
Trie primary factors influencing denting are degree of superheat and bulk water chloride
and oxygen concentrations. Chlorides result in an acidic secondary water chemistry
environment, which causes rapid corrosion of the carbon steel support plate when sufficient
oxygen is present. Copper oxide may also play an important role as a supplier of oxygen to
the carbon steel support plates. Sulphates (e.g., from condensate polisher leakage) are believed
to cause denting in the same manner as chlorides, though the laboratory test database is not
as extensive. Denting was relatively uncommon when most plants used phosphate water
chemistry, since the phosphates kept the crevice pH high. Although modifications and
attention to secondary-side water chemistry have reduced denting to a lesser concern, denting
is still considered a degradation concern, particularly if a unit (a) has experienced one or more
major secondary-side intrusions of contaminants, or (b) is constructed with low-temperature
mill-annealed tubing and is, therefore, susceptible to PWSCC even at small-size dents [EPRI
1985, Theus and Daniel 1984, Frank 1984, Nordmann et al. 1983].
Denting of Alloy 600 tubes at tube-to-tube support plate intersections was first identified
as a significant steam generator degradation mechanism in about 1975, shortly after the time
when many PWRs switched from phosphate to AVT secondary-side water chemistry, and this
degradation mechanism became the primary cause of steam generator tube plugging during
the period 1976 through 1979 (Fig. 13). As of December 1993, 1471 RSG tubes at 41 plants
(four Combustion Engineering and 37 Westinghouse-type plants) had been plugged because
of tubesheet and sludge pile denting and 9092 RSG tubes at 17 plants (four Combustion
Engineering and 13 Westinghouse-type plants) had been plugged because of support plate
denting (EPRI 1994). Significant support plate denting occurred at only five plants: Millstone
Unit 2 (796 tubes), Surry Unit 1 (1996 tubes), Surry Unit 2 (1964 tubes), Turkey Point Unit
3 (1249 tubes), and Turkey Point Unit 4 (1835 tubes); all of the original steam generators at
those plants have since been replaced. The majority of the support plate denting has occurred
on the hot leg side at plants with seawater or brackish water for condenser cooling.
CANDU units with the older Alloy 600 and Monel 400 steam generators with carbon
steel supports have also experienced tube deformation due to deposit buildup in the tube-
support gaps and corrosion of the supports. However, tube cracking has not been detected in
the deformed areas.
4.1.6. High-Cycle Fatigue
The combination of high vibration amplitude and low fatigue strength may lead to
catastrophic fatigue failure. Vibration occurs in steam generators with high recirculation flow
42
factors (causing flow-induced vibrations in the U-bend region) and improper antivibration bar
support. A high mean stress (e.g., residual stress) significantly reduces the fatigue strength.
Therefore, tubes with dents at the top tube support plate in the U-bend region of the RSGs
are susceptible to high-cycle fatigue failure.
High-cycle fatigue failures have occurred in the U-bend regions of the North Anna Unit
1 and Mharna Unit 2 steam generators. Though high-cycle fatigue from tube vibrations is not
a general problem in PWR steam generators, tube ruptures, such as those at North Anna and
Mhama Unit 2, are of particular concern because they were 360° breaks located high up in
the steam generator where the leak location can more readily become uncovered by secondary
water. This can allow escape of fission products from the primary coolant without partitioning
in the secondary water. For example, upon failure of the Mihama Unit 2 Steam Generator A
tube, the primary system leak rate rapidly escalated from a very low level to a value
exceeding the normal capacity of the charging pumps. The ruptured tube eventually released
about 55 000 kg (55 tonnes) of primary coolant to the secondary-coolant system.
Approximately 1300 kg (1.3 tonnes) of steam, 0.6 curie of radioactive noble gases, and 0.01
curie of radioactive iodine subsequently escaped from the damaged steam generator's relief
valve to the environment. The reactor core remained submerged owing to the operation of the
high-pressure injection system.
Most of the earlier tube failures in CANDU steam generators tubed with Alloy 600 have
been due to high cycle fatigue. These failures were initiated at either fret marks or more
recently at stress corrosion cracks and are caused by flow-induced vibration at the U-bend
area and at the seventh support plate. These failures continue to occur in the older CANDU
steam generators.
4.1.7. Wastage
Phosphate wastage was the major cause of tube failures in PWR steam generators until
about 1976. However it is no longer an active degradation mechanism in most of the PWRs
because phosphate water chemistry is no longer used.
4.2. PWR ONCE-THROUGH STEAM GENERATOR TUBES
Once-through steam generators in the USA use the same Alloy 600 tubing materials as
RSGs, yet these steam generators have experienced substantially fewer tube failures. The
lower failure rate is attributed to the differences in the steam generator design, manufacturing
processes, and operation. Many of the chemical concentration processes do not operate in
once-through steam generators, as they do in RSGs [Turner 1988]. Table VIII lists
once-through steam generator tube degradation mechanisms, sites, stressors, failure mode and
inspection methods. The most common tube degradation mechanisms are briefly discussed
here. However, as noted below, even these mechanisms affect a very small percentage of the
tubes in service.
43
TABLE VIE. SUMMARY OF ONCE-THROUGH STEAM GENERATOR TUBE
DEGRADATION PROCESSES
Degradation site(s) Degradation Potential failure In-service inspection

Rank1 Stressors rnechanism(s) mode methods)
1 Outside surfaces of the Velocities, sizes, shapes, Erosion-corrosion Wear of material Eddy-current testing
tubes on the periphery impact angle, and hardness from impingement
of the tube bundle near of particles of particles
the 14th tube support
plate
Tube outside surfaces Aggressive chemicals, Environmentally Circumferential Eddy-current testing
near die upper tubesheet vibration assisted high-cycle cracks
and the open lane or fatigue
near the uppermost tube
support plate and the
open lane
Inside surfaces of tubes Sodium thiosulphate, air Low-temperature Circumferential Eddy-current testing
near the upper tubesheet primary-side stress cracks
roll transitions and corrosion cracking
welds (primary side)
Based on operating experience and number of defects
4.2.1. Erosion-Corrosion
Erosion-corrosion results from entrained solid particles impinging on metal surfaces

causing material removal, wear, and mechanical damage, especially if there is a protective
surface film present, hi a corrosive environment, the erosion process first removes the
protective film from the tube, thus making the tube susceptible to more corrosion and then
more erosion. Inspection of removed tubes indicates that erosion-corrosion has occurred in
once-through steam generators on the outside of the heat exchanger tubes, principally around
the fourteenth tube support plate at the periphery of the tube bundle.
The fraction of tubes for all once-through steam generators affected by erosion-corrosion
is small. Through December 1993, 1622 tubes (about 0.75% of the tubes in service) have
been taken out of service due to erosion-corrosion. More than half (991) have been from one
plant, hence, the mechanism is not occurring at the same rate in all steam generators.
4.2.2. Hgh-Cycle Fatigue
Through-wall circumferential cracking has occurred in once-through steam generator

tubes at the top tube support plate (i.e., 15th tube support plate) and at the bottom of the
upper tubesheet in the inspection lane region. The inspection lane region includes about three
rows of tubes on either side of the inspection lane and a few additional rows at the periphery.
The cracks initiated at the outside diameter of the tubes and propagated circumferentially in
a transgranular mode. Tube samples revealed a serpentine band of metal loss in the areas near
the upper tube support plate and just below the lower face of the upper tubesheet (EPRI 1985,
44
Theus and Daniel 1984). Sometimes these metal loss areas contained microcracks that acted
as the site for fatigue crack initiation. Laboratory tests indicate that the corrosive metal loss,
including the microcracks, can be achieved with concentrations of sodium sulphate, silicates,
and chlorides (Monter and Theus 1982). Thus, the degradation mechanism has been described
as environmentally assisted high-cycle fatigue.
The stressors for this corrosion fatigue cracking are believed to be deposits of
concentrated impurities and cyclic vibration. Evaporation of the secondary-side water in the
lower elevations of the once-through steam generators concentrates any contaminants or
impurities into the remaining droplets. The steam flow then carries these droplets up the open
inspection lane to the upper tubesheet area, where the droplets impinge on the hot tubes
around the inspection lane, dry out, and deposit the impurities. This process further
concentrates the chemicals at selected locations on the steam generator tubes. Through
December 1993, a total of 106 tubes (0.05% of the tubes in service) at six plants have been
taken out of service due to this mechanism.
4.2.3. Low-Temperature Primary-Side Stress Corrosion Cracking
Stress corrosion cracking on the inside surfaces (primary side) was detected in the tubes
of a once-through steam generator at IM-l in 1981, where essentially all the tubes were
affected and 1619 tubes plugged and 502 tubes sleeved [Jones et al. 1982, Giacobbe et al.
1988]. The incident is unique because the combination of conditions necessary to promote this
type of attack is not expected to occur at other plants. Partially reduced sulphur species (e.g.,
sodium thiosulphate) had inadvertently been introduced into the primary system from the
containment spray system It is believed that aggressive concentrations of sodium thiosulphate
and oxidizing conditions developed in the failure area from dryout and exposure to air. Most
of the defects were circumferential in geometry and located in the upper part of the upper
tubesheet near the weld heat-affected zone or the roll transition. The main protection against
recurrence of this type of incident in once-through steam generators with sensitized tubing
(due to heat treatment, see Section 3.2.1) is to avoid acidic oxidizing conditions by strict
water chemistry controls and proper lay-up using strict procedure controls.
4.2.4. Outside Diameter Inteigranular Stress Corrosion Cracking (IGSGQ and

Inteigramilar Attack (IGA)
As discussed in Section 4.1.2 above, IGSCC requires tensile stress, material susceptibility
and a corrosive environment. IGSCC cracks occur along the grain boundaries, normal to the
maximum principle stress. IGA is characterized by local, corrosive loss of material along the
grain boundaries. Both mechanisms require a concentration of corrosive impurities on the
outside surface of the tubing.
Through December 1993, 543 tubes (about 0.25% of the once-through steam generator
tubes in service) at four plants were removed from service or repaired due to IGSCC/IGA.
The damage primarily occurred near the upper tubesheet (492 tubes).
45
4.3. WWER STEAM GENERATOR TUBES
The horizontal, U-shaped tubing used in the WWER-440 and WWER-1000 reactors has
been relatively trouble free. The WWER tubing is made of titanium-stabilized austenitic
stainless steel with about 0.08% carbon, 18% chrome, 10% nickel, < 1% titanium and the rest
mostly iron. Through the end of 1989, only about 2815 WWER-440 tubes and 655 WWER-
1000 tubes had been plugged out of a total of 1 774 480 tubes in operation; e.g., only about
0.2% of the total number of tubes had been plugged (Titov et al. 1992).
The main cause of damage has been outside surface stress corrosion cracking due to poor
secondary side water chemistry, primarily chloride ion and oxygen excursions, but also low
pH and the presence of various organic compounds. Secondary side chloride ion
concentrations of several hundred to several thousand ppb have been reported for relatively
significant times (Rassokhin et al. 1992). Also, the effects of the chloride ions on the stress
corrosion rate have been accelerated due to the presence of porous crud deposits in quantities
in excess of 150 g/m2 (the recommended limit). The chloride ions tend to concentrate in the
crud capillary structures by factors of 105 to 106 (Titov et al. 1992, Mamet and Martynova
1993). At some WWER plants, the pH has dropped below 7.8 (the original lower limit which
has now been revised to 8.8 for the feedwater and 8.0 for the blowdown water as listed in
Table VI) for up to 20% of the overall running time, and to the range of 5-6 for up to 2%
of the running time. Also, up to 700 ppb of acetic acid (due to organic compound breakdown)
has been found in the feedwater at several plants (Martynova and Mamet, 1991).
These secondary side chemistry excursions have also caused pitting corrosion, for
example at the grid spacer locations at the Novovoronezh Units. There have also been a few
collector weld defects which have resulted in plugged tubes. The repair criteria for the
WWER steam generators is tube leakage and the method is plugging; e.g., leaking tubes are
plugged, other indications (part throughwall defects) are usually ignored.
4.4. TUBE RUPTURE EVENTS
4.4.1. Tube Ruptures
The leak rate, degradation mechanism, rupture size, rupture location, and stressor and
contributing factor information associated with ten steam generator tube rupture events is
surnmarized in Table DC (MacDonald et al. 1996). These ruptures have occurred over the last
20 years at a rate of about one every 2-3 years and may continue to occur. The maximum
leak rates have ranged from 470 L/min (125 gal/min) to 2880 L/min (760 gal/min). Maximum
leak rates less than about 380 L/min (100 gal/min) are considered by the USNRC to be from
tube defects rather than tube ruptures, since that amount of leakage is within the normal
capacity of the charging systems. The highest possible leak rates calculated for a single tube
rupture are on the order of 3800 L/min (1000 gal/min).
Five different tube degradation mechanisms caused the ten ruptures: three ruptures were
caused by ODSCC, two ruptures were caused by high-cycle fatigue, two ruptures were caused
by loose parts wear, two ruptures were caused by PWSCC, and one rupture was caused by
wastage.
46
TABLE IX SUMMARY OF THE LEAK RATE, DEGRADATION MECHANISM, RUPTURE SIZE, RUPTLJRE LOCATION, AND STRESSOR INFORMATION
ASSOCIATED WITH TEN STEAM GENERATOR TUBE RUPTURES
Maximum
Plant, Leak Rate Degradation Rupture Rupture Stressors and contributing factors
02/26/75 Point Beach- 1 125 Wastage 2 adjacent ruptured bulges Slightly above the tubesheet, outer row on Large sludge pile, ineffective cleaning
W-44 each about 20 mm long and the hot leg side
wide
09/15/76 Suny-2 330*" PWSCC 1 14 3 mm long axial crack Top of U-bend (apex) in Row 1, Column 7 High stresses and ovalizahon caused by inward
W-51 movement of the legs due to support plate
deformation
06/25/79 Doel-2 135 PWSCC 100 mm long axial crack Top of the U-bend in Row 1, Column 24 ligh residual stresses due to ovalizahon during
ACE-44 abncahon
10/02/79 Prauie Is -1 336<" Loose Parts 38 mm long axial fishmouth Tube bundle outer surface, 76 mm above Sludge lancing equipment left in the steam
W-51 Wear opening the tubesheet on the hot leg side, Row 4, generator
Column 1
01/25/82 Ginna 760"' Loose Parts 100 mm long axial fishmouth 127 mm above the tubesheet on the hot leg Loose parts (baffle plate debns) left in the
W-44 Wear, Fretting opening side, Row 42, Column 55 (third row in steam generator, wear of peripheral tubes,
from the bundle periphery) fretting of inner tubes
05/16/84 Fort Calhoun 112 ODSCC 32 mm long axial crack (small Horizontal run at the top, between the Tube deformation caused by corrosion of the
CE fishmouth opening) vertical barwing support bars on the hot vertical batwing support bars, caustic impurities
leg side, Row 84, Column 29, the rupture on the secondary side
faced down
07/15/87 Morth Anna-1 637 High-Cycle 360° circumferential break Fop of the 7th tube support plate on the High-cycle vibration, denting, lack of AVB
W-51 Fatigue cold leg side, Row 9, Column 51 upport
03/07/89 VfcGuire-1 500 ODSCC 95 mm long axial crack in a 711 mm above the tubesheet at the lower jong shallow groove, possibly a contaminant
W-D2 645 mm long groove, 9 5 mm tube support plate on the cold leg side,
wide at the maximum point Row 18, Column 25
02/09/91 VIihama-2 *500 High-Cycle 360° circumferential break Fop of the 6th (upper) tube support plate iigh-cycle vibration, lack of AVB support
MHI-44 Fatigue on the cold leg side, Row 14, Column 45
03/14/93 Palo Verde-2 240 ODSCC 65 mm long axial fishmouth Freespan region between the 08H and 09H Tube-to-tube crevice formation, bridging
CE-80 opening in a 250 mm long tube support structures on the hot leg side, deposits, caustic secondary water chemistry,
axial crack Row 117, Column 144 susceptible material
(I)
NRC estimates
Additional ruptures caused by wastage are unlikely because only three reactors worldwide are
now using phosphate water chemistry.
Additional ruptures due to high-cycle fatigue in Westinghouse-type steam generators are

less likely than a few years ago because most operators have inspected their steam generators
to assure that the AVBs are properly placed and new steam generators are being more
carefully fabricated with more and better AVBs. However, the Indian Point Unit 3 experience
suggests that such failures are possible even with proper AVB support. [A 456 L/h (120
gal/h) leak developed at Indian Point Unit 3 on 19 October 1988. Subsequent inspections
identified a 250° circumferential high-cycle fatigue crack in the tube in Row 45, Column 51,
just above the upper most support plate. The tube was dented at the support plate due to
support plate corrosion, however, the Indian Point Unit 3 tube was properly supported by its
AVBs.] Loose parts and other foreign objects continue to be left in some steam generators
and additional ruptures of tubes due to loose parts wear are possible. Also, extensive primary
water and outside diameter stress corrosion cracking has occurred in certain steam generators
and more ruptures caused by those mechanisms are possible.
The rupture locations have generally been either just above the tubesheet (three ruptures),
or in the U-bend region (six ruptures). Only the McGuire rupture was near one of the lower
support plates. The ruptures caused by loose parts wear have occurred just above the
tubesheet whereas the ruptures caused by high-cycle fatigue have occurred just above the top
tube support plate. Any future ruptures caused by those mechanisms will probably occur in
the same locations.
The three ruptures caused by ODSCC appear to each have some unique contributing
factors. The Fort Calhoun tube was subjected to high stresses caused by corrosion of the
vertical batwing support bars. The McGuire rupture was located in a long shallow groove
which was probably created during fabrication. The Palo Verde rupture occurred in a tube
with a susceptible (and abnormal) nricrostructure. However, excessive caustic impurities on
the secondary side were part of the problem in all three cases.
The plant transient information is siirnmarized in Table X (MacDonald et al, 1996). The
operators were expected to (a) maintain the primary coolant subcooled, (b) minimize the
leakage from the reactor coolant system to the defective steam generator secondary side, and
(c) minimize the release of radioactive material from the damaged steam generator. Timing
is critical to the successful management of a steam generator tube rupture event. The key
operator actions that must be accomplished in a timely manner include:
• Recognition that a steam generator tube rupture event is occurring.

• Control of the pressurizer level using the charging pumps and letdown line (if the
rupture is small).
• Power reduction/trip.
• Isolation of the defective steam generator.
Reactor coolant system cooldown including pumped flow to the intact steam generators
and intact steam generator steam dumps to the condenser or atmosphere.
• Reactor coolant system depressurization which generally requires throttling the safety
injection and use of the pressurizer sprays or PORVs.
48
TABLE X SUMMARY OP PLANT TRANSIENI INIORMATION
Point Beach l Surry2 Doel2 Frame Is 1 Ginna Fort Calhoun North Anna 1 McGuire 1 Mihama 2 Palo Verde 2
Maximum leak rale (gpm) 125 330 135 336 760 112 637 500 500 240
At power Yes Yes No Yes Yes No Yes Yes Yes Yes
First indication of rupture Air ejector Pressure air ejector rad Pressure Air ejector Air ejector Pressure Main sleam line rad Mam sleam line rad Air ejector Pressure MSL rad
rad rad rad rad
Time operators recognized SGTR 24 28 mm < S mm «9 mm 5 1 8 5 mm < 1 mm •32 mm < 5 mm < 1 mm *5 mm < 57 mm
Second, third charging pumps started (mm) 2 19 5 (2nd) 1 8 < 15 9 10 1 25 0 0 increased flow 18 nun 4 (2nd) 5 (3rd) 2 (3rd)
Letdown line isolated 8 mm S mm 2 4 mm 3 mm 24 mm 3 mm S (reduced) 6 mm
Load reduction started 30 mm 7 mm N/A 7 mm 1 S mm N/A 3 mm 4 mm 7 mm No
Manual reactor mp 47 mm (al 25% power) 10 mm (al 70% power) N/A No No N/A 5 nun 8/9 mm No 13 mm
Automatic reactor scram No No N/A 10 15 mm 3 mm N/A No No 10 mm No
Automatic safety injection No (blocked al 54 mm) No manual SI at 1 1 mm 192 mm 1023 mm 3 mm No 5 3 mm No (blocked at 23 mm) 10 1 mm 13 2 mm
Defective steam generator isolated 5! mln II mln 94 mln 17 mln IS mln 40 mln 11 mln 11 mln 21 mln 1 h 54 mln
mam steam isolation valve closed 48mm 18 mm closed 27 nun 15 mm 40 mm 16 mm *l 1 mm 22 mm 2 h 54 mm
mam feedwaler valve closed 58 mm 1 1 mm closed 10 IS mm 3 mm closed 5 mm *9 nun 12 mm 13 mm
auxiliary feedwaler flow None II mm 41 SO nun 3 7mm 0 32 mm 53 16 mm 12 mm
feed to Terry turbine No No 4 7 mm 0 18 nun 8 mm
safely valve open No No No No 54 63 114 132 mm No No
atmospheric dump valves open No No No 1 2 sec at scram No No 39 49 59 mm
RCS cooldown started SI mln 21 mln 41 mln 1 h 36 mln 2 mln 41 mln 19 mln 14 mln 3 5 h 22 mln 2 h, 47 mln
RCP on defective sleam generator Inpped 66 mm II Sh 174 mm 12 mm 4 mm 42 nun 43 mm 46 mm 13 mm (one)
RCP on intact sleam generators tnpped No 19 mm (One) 13 mm 4 mm 43 mm 43 nun (one) 47 mm 13 mm (one)
RCP on intact steam generators restarted N/A No 7 hours 116 mm 1 h 17 mm
mlacl sleam generator steam dumps SI mm Yes 2 3 2 5 75 mm 41 mm 19 mm 14 25 mm 3 S h 22 37 mm
62 94 mm
RCS depressunzalion started SI mln 16 mln 68 88 mln 42 mln 2, 73 mill 10 mln 16 mln 14 mln 54 mln 1 h 30 mln
safely injection throttled Yes (61 mm) Yes (16 mm) 2 h 47 mm No N/A
safely injection stopped N/A 16 mm* 62 mm 42 52 mm 73 mm N/A 57 mm (two)
pressunzcr spray used 1 h 40 mm Yes 28 mm No No Yes (72 mm) Yes 54 mm*
pressunzer PORV open No No blocked Yes (43 mm) 42 44 mm No 34 mm No (tned)
charging pumps slopped 73 79 nun 21 mm 3 mm 27 mm (2) 37mm(l)
Reactor coolant system and defective steam generator secondary «7h - 1h ' 61 mm 3 h 2 mm •> 34 mm 47 mm 10 h 37 mm 1 h 8 mm 3to4h
pressure equal
R1IR m operation 3 h 5 mm - II Sh 3 h IS mm 16 h 26 mm 21 h 3 5 m m 3 h 47 mm 5 h 49 mm 17 h 8Sh 6h
Excessive level in defected sleam generator Yes No No No Yes tcs No No No No
b The auiiliary spray wai started u 54 minutes stopped at I hour 8 minutes and then set partially open at I hour 19 minutes
It should be noted that based on the training that reactor operators receive prior to
licensing, a steam generator tube rupture is normally easily recognizable. The operators have
several indicators that can be referred to that point to the fact that a tube rupture is occurring.
The steam line radiation monitors and the air ejector radiation monitors are the prime
indicators and are used as Emergency Operating Procedure entry conditions. The operating
procedures that are utilized to combat the transient assume that the plant is at power and that
the systems are aligned properly.
The success of the operators, as indicated by the times these activities started or finished
in Table IX, is mixed. For example, the Point Beach, Fort Calhoun, and Palo Verde operators
took a relatively long time (up to 28, 32, and 57 minutes, respectively) to realize (or prove
to themselves) that a steam generator tube rupture had occurred. The result was that they were
slow to start reducing power (30 minutes at Point Beach where the maximum leak rate was
relatively low, 13 minutes at Palo Verde where the leak rate and the initial pressure drop were
larger) and slow to isolate the defective steam generators (58, 40, and 174 minutes,
respectively). By contrast, the Ginna, North Anna, McGuire, Surry, and Mhama operators
recognized that a steam generator tube rupture event was happening within a few minutes of
the first alarm. Their load reductions started within 1.5, 3, 4, 7 and 7 minutes, respectively,
and their defective steam generators were isolated within 15, 18, 11, 18, and 22 minutes,
respectively. (The defective steam generator at Mhama would have been isolated at 15
minutes, had the main steam line isolation valve worked properly.) It should be noted that
it is harder for the operators of a plant at or near hot standby (Doel and Fort Calhoun) to
detect a steam generator tube rupture. But the operators at Point Beach, Palo Verde, and
probably Prairie Island, should have been able to recognize and identify the event much
faster.
Also, a significant drop in pressurizer level should signal the operators to start and set
the second and third charging pumps at full flow as well as reduce or isolate the letdown
flow, and that happened in most cases. However, the third charging pump did not start at
Doel until about 15 minutes, the second and third charging pumps did not start at Prairie
Island until 9 and 10 minutes, and the charging pumps at Fort Calhoun were not at full flow
until 18 minutes. Adequate charging flow can prevent safety injection (for smaller ruptures)
and allow the pressurizer to be used to help control the early depressurization. It is realized
that in some instances the third charging pump may be a low volume, high discharge pressure
pump that is normally utilized for make-up and is ineffectual in supplying large quantities of
water in an emergency.
Another area where timely actions were important is the cooldown and depressurization
of the primary system. It is very important to get the reactor coolant system pressure down
to a value below the defective steam generator secondary side pressure and keep it there
(slightly below, but not so far below that the backflow will significantly effect the primary
system boron concentration) while at the same time keeping the reactor coolant system fully
subcooled Reactor coolant system pressures above the defective steam generator secondary
side pressure for long periods of time result in overfill of the steam generator secondary side
and unnecessary radioactive material releases to the environment. The North Anna, Surry,
Prairie Island, and Mhama reactor coolant system pressures were reduced to their defective
steam generator secondary pressures in 34, 60, 61, and 68 minutes, respectively, and there
50
TABLE XI. RECENT INCIPIENT STEAM GENERATOR TUBE RUPTURE EVENTS IN THE USA
Plant Date Maximum Leak Defect Size Defect Location Degradation

Rate Mechanism
Braidwood Unit 1 23 October 1993 *47 LA (12.5 gal/h) 330 mm (1.3 inch) crack Above the top tube support plate near AVB ODSCC
Arkansas Nuclear One, 9 March 1992 57L/h(15gal/h) Circumferential Hot leg side of the tube in Row 67, Column 109, ODSCC
Unit 2 throughwall crack 4.8 mm above the tubesheet in the explosive
transition region
McQuire Unit 1 16 January 1992 37 L/h (10 gal/h) 250 mm long axial crack Cold leg side of the tube in Row 47, Column 46, ODSCC
130 mm above the lower tube support plate
Maine Yankee 17 December 1990 318 L/h (84 gal/h) 100 mm long axial crack Top of the U-bend of the tube in Row 6, Column ODSCC
43
Three Mile Island Unit 1 6 March 1990 «1 15 L/h (30 gal/h) 360° circumferential crack Peripheral tube A77-1 next to the open inspection High-cycle fatigue
lane, bottom of upper tubesheet (environmentally
assisted)
Beaver Valley Unit 2 21 June 1989 80 Uh (21 gal/h) 97% throughwall wear, Hot leg side of the tube in Row 31, Column 16, Loose parts
small rupture 25 mm above the tubesheet damage
Indian Point Unit 3 19 October 1988 456 17h (120 gal/h) 250° circumferential crack Tube in Row 45, Column 51, just above upper High-cycle fatigue,
support plate denting
were no defective steam generator overfill problems. The Point Beach, Ginna, and Fort
Calhoun, reactor coolant system pressures were held well above the defective steam generator
secondary side pressures for considerably longer times (about 7, 3, and an unknown number
of hours, respectively) and the defective steam generators overfilled. The McGuire
depressurization also took a very long time (10 hours, 47 minutes), but the defective steam
generator at McGuire was not overfilled because of releases to the condenser and through the
condenser vent, to atmosphere.
Despite these variations in timing, it should also be noted that in all cases the plants
were properly cooled down and the radioactive material releases were small and well below
regulatory limits. Also the operator performance was sometimes hampered by inadequate
Emergency Operating Procedures (Palo Verde, for example) or by defective equipment
(Mihama, for example). At other times the operator was hampered by plant conditions that
did not allow rapid employment of Emergency Operating Procedures. There are still numerous
reasons for (a) continued operator training on steam generator tube ruptures and (b) training
on the recognition of events based on the indications that are available. Neither training
method should be utilized by itself. It appears that the majority of actions that were carried
out were accomplished in accordance with the published procedures. Deviations from
procedures appeared to be thought out in advance.
4.4.2. Incipient Tube Rupture Events
Seven incipient tube rupture events which occurred in the USA during the last seven
years are summarized in Table XI (MacDonald et al. 1996).
For some of these incipient steam generator tube rupture events, the operators were able
to quickly shut down the reactor and isolate the detective steam generator. In other cases, the
cracks stopped growing for unknown reasons. These actions limited the contamination of the
secondary coolant and may have prevented actual tube rupture. Also, some of these events
demonstrated how quickly very low leak rates can increase as the crack grows. Leak rate
monitoring programmes that provide close to real time information can limit the frequency
of steam generator tube ruptures. "At some sites, data from the air ejector radiation monitors
is continuously displayed in the control room. At other sites, main steamline radiation
monitors promptly detect increases in nitrogen-16 activity. When combined with appropriate
alarm setpoints and operational limits, this information can quickly alert operators to
implement response procedures to monitor increases in leak rates or to shut down the reactor
and isolate the affected steam generator" (USNRC 1994).
4.5 PWR AND CANDU STEAM GENERATOR SHELL, FEEDWATER NOZZLE AND
TUBESHEET
This section discusses degradation mechanisms in steam generator shells and feedwater
nozzles. Corrosion fatigue, high-cycle thermal fatigue, and stress corrosion cracking have
caused cracking on the secondary sides of the PWR steam generator shells. PWR primary side
degradation has not been observed and there has been no PWR tubesheet or CANDU shell,
nozzle, or tubesheet degradation reported. Thermal fatigue and erosion-corrosion are
52
TABLE XII. SUMMARY OF DEGRADATION PROCESSES FOR PWR FEEDWATER NOZZLES
AND STEAM GENERATOR SHELL
In-service
Degradation Potential failure inspection
Rank1 Degradation site(s) Stressors mechamsm(s) mode method(s)
1 Feedwater nozzle and Flow velocity, O2 content High and low cycle Rupture from wall Ultrasonic testing
nozzle-to-piping weld and pH level in feedwater, fatigue, erosion- thinning, leakage radiography
impurities, stratified flows, corrosion. through fatigue
thermal shocks, water cracks, rupture
hammer, plant transients from water
hammer
2 Steam generator shell Plant transients, oxygenated Corrosion-fatigue, Leakage through Ultrasonic testing,
girth welds coolant containing copper stress corrosion fatigue or stress- radiography
oxide, m-leakage of cracking corrosion cracks
brackish water through
condenser tubes, residual
stresses
Feedwater nozzle bore, Leakage of feedwater High-cycle thermal Leakage through Ultrasonic testing,
blend radius, shell inside through the nozzle thermal fatigue fatigue cracks radiography
surface beneath the sleeve joint causing
nozzle turbulent mixing of cold
feedwater and hot steam
generator coolant
4 J-tubes and feedrmg Flow velocity, 0, content Erosion-corrosion Damage caused Problem remedied,
and pH level in feedwater, by loose parts, inspection
impurities thermal fatigue to unnecessary
shell
•Currently performed but not included in the in-service inspection requirements
responsible for most of the ageing degradation that has occurred in PWR feedwater nozzles
and the nozzle-to-pipe weld regions. Ageing degradation may cause leakage but probably not
failure, however, it may also so weaken the system and reduce the safety margin that another
event, such as a pressure pulse or a water hammer, could be the final cause of a rupture.
Primary side divider plate damage has occurred at some CANDU units.
Table XII lists and ranks by importance the degradation mechanisms, sites, stressors,
failure modes, consequences, and inspection methods for the PWR feedwater nozzles and the
steam generator shells. The feedwater nozzle is ranked highest, because a break at this point
would cause a much larger leak than from a steam generator shell crack and might not be
isolated from the steam generator, thus leading to rapid blowdown of the steam generator.
Such a break would challenge the integrity of any severely degraded tubes.
4.5.1. Corrosion-Fatigue
Steam Generator Shells. High-amplitude, low-frequency cyclic stresses combined with

coolant containing oxygen and copper oxides have caused corrosion-fatigue damage to the
upper girth weld, i.e. upper shell to transition cone weld, in about seven recirculating steam
generator shells in the USA (Note that significant concentrations of copper oxides are
associated with copper alloys in the feedtrain.) The presence of oxygen and copper oxides
53
probably contributes to the formation of surface pits, which act as stress raisers, and therefore,
as sites for fatigue crack initiation in the steam generator shell. During a few transient events,
the water level in the steam generator drops below the girth weld region, and the incoming
feedwater impinges on the girth weld and produces rather high stresses (Bamford, Rao, and
Houtman 1992). Also, fluctuations in the steam generator water level will impose thermal
fatigue cycles on the steam generator shell. Circumferential cracks have been observed in the
girth weld under the feedwater nozzle, mainly in the heat-affected zone, with little penetration
in the base metal. This suggests that the fracture toughness of the heat-affected zone was
substantially lower than that of the base metal, and that the stresses were large enough to
drive the cracks through the heat-affected zone but not through the base metal [Kobayashi and
Shockey 1991].
Feedwater Nozzles. Corrosion-fatigue cracks, caused by coolant thermal stratification and

the stress concentrations at a counterbore (a joint between the feedwater nozzle and piping
with a geometric discontinuity), have been observed in the vicinity of the feedwater nozzles.
Under low feedwater flow conditions, typically during hot standby when the feedwater is
supplied by the auxiliary feedwater system, the relatively cool feedwater tends to flow along
the bottom of the horizontal sections of the piping adjacent to the feedwater nozzle, with the
top portion containing hot water. This thermal stratification may lead to two different stressors
which cause fatigue damage: cyclic local stratification and "thermal striping." Cyclic local
stratification stresses, caused by small auxiliary feedwater flow fluctuations and subsequent
changes in elevation of the interface between the hot and cold layers, can produce significant
stress changes at a point in the pipe cross-section. Thermal striping, due to turbulent mixing
at the interface of the hot and cold layers, can produce high cycle fatigue crack initiation,
generally a surface effect. Thermal striping does not propagate cracks; however, cyclic
thermal stratification may propagate shallow cracks caused by thermal striping. The stress
concentration at the sharp transition from the smaller thickness nozzle to the larger thickness
feedwater pipe near the nozzle/pipe weld counterbore can also promote cracking in this
region, [Cofie et al., 1994].
On 25 June 1979, the USNRC issued Bulletin 79-13 requesting examinations of the
feedwater nozzles and adjacent piping in the USA to address the safety concerns raised by
fatigue cracking [USNRC 1979]. The resulting inspections revealed pipe cracks in the vicinity
of the feedwater nozzles at 18 of the 54 facilities inspected [Cofie et al., 1994]. All cracks
were corrosion-fatigue cracks caused by cyclic thermal stratification, except the cracks at one
plant, which were identified as stress corrosion cracking [USNRC 1979b]. Recently, feedwater
fatigue cracking has again been observed at several US plants, including a through-wall crack
at one unit. This cracking appears to have been caused by high stresses at the counterbore and
fluctuations in the auxiliary feed water flow, water chemistry may also have played a
secondary role.
Both carbon steel piping material and low-alloy steam generator shell material are
susceptible to corrosion fatigue if they contain sulphur inclusions, such as manganese
sulphides [Bamford et al. 1987, Van der Sluys and Cullen 1987]. The morphology and
distribution of the sulphides can cause the crack growth in low-alloy pressure vessel steels
to differ by a factor of two, depending on the crack plane orientation [Van der Sluys 1982].
Environmental effects appear highest for steels with medium-to-high concentrations of sulphur
54
(>0.015 wt%) in highly oxygenated water; environmental effects may be negligible in low
sulphur (O.010 wt%) steels in deoxygenated water.
4.5.2. Transgranular Stress Corrosion Cracking
Steam generator shell material subjected to high tensile stresses and oxygenated
secondary coolant containing copper oxides is susceptible to transgranular stress corrosion
cracking. High tensile stresses include both weld residual and operating stresses.
Transgranular stress corrosion cracking and corrosion fatigue are differentiated by their load
histories. Transgranular stress corrosion cracking occurs when the applied stresses are constant
or have a very small fluctuation, i.e., the ratio of the minimurr^to-maximurn stress intensity
factors is close to one. Corrosion fatigue occurs when the applied stresses are cyclic and the
ratio of stress intensity factors is smaller than about 0.95.
Circumferential cracks and linear indications have been detected on the inside surface
of the girth welds in 18 steam generators in the USA, all of which are Westinghouse Models
44 and 51 with a feedring design [USNRC 1990a]. This type of cracking was first observed
in 1982 when a girth weld of a steam generator leaked at a US plant [USNRC 1982]. Linear
indications have also been detected at least one non-US plant. In most of these cases, the girth
weld region was predominantly subject to static loads and the cracking was caused by
transgranular stress corrosion cracking.
Leak-before-break analyses show that a stress corrosion crack will grow through the shell
wall and produce a measurable leak before it exceeds the critical flaw size and the vessel
ruptures [Westinghouse 1990]. Field experience to date supports this analysis. Inspection port
holes in the steam generators have also experienced cracking, most likely stress corrosion
cracking, on the inside surface. Grinding of the inspection port hole might have introduced
the residual stresses needed for stress corrosion cracking.
4.5.3. High-Cycle Fatigue
High-cycle fatigue degradation can be caused by cyclic thermal stratification, thermal

striping, and turbulent mixing of leaking cold feedwater (if any) with hot steam generator
coolant (see Section 4.5.1). Any leakage of the feedwater through the feedwater
nozzle-thermal sleeve joint can cause thermal stratification, turbulent mixing, and thermal
shocks in the feedwater nozzle. These thermal stresses can promote fatigue damage in the
nozzle bore, nozzle blend radius, and the inside surface of the shell. At one PWR plant in the
USA, the feedwater nozzle bore region, blend radius, steam generator shell inside surface
beneath the nozzle (see Fig. 11), and feedring support bracket welds have all experienced
cracking, probably due to both thermal fatigue and stress corrosion.
4.5.4. Erosion-Corrosion
Erosion-corrosion is a flow-assisted corrosion mechanism that affects carbon steel piping

carrying single-phase, subcooled feedwater and steamlines carrying wet steam. The damage
caused by erosion-corrosion is higher than damage attributed to erosion or corrosion alone.
Carbon steel feedwater piping corrodes during normal operation, forming a thin layer of iron
55
Oxide-water
interface (E)
Metal-oxide
interface
nmmMmmmmmmlwmum__ (A*B)
A. Iron hydroxides are generated: Fe + 2Hg O —o- Fe(OH)2 + H2
B. Magnetite is formed according to the Schikorr reaction:
3Fe(OH)2 —e» Fe3 O4 + HZ + 2H2 O
C. A fraction of the hydroxides formed in step B and hydrogen generated in steps A and
B diffuse along pores in the oxide
D. Magnetite can dissolve in the pores
E. Magnetite dissolves at the oxide-water interface
F. Water flow removes the dissolved species by a convection mass transfer mechanism
G. Solid particles break off porous oxide layer by a mechanical
erosion mechanism
FIG. 18. Phenomena occurring during erosion-corrosion (Sanchez-Cddera 1984).
oxide [mostly magnetite (Fe^)] on the inside surface. This layer protects the underlying
piping material from the corrosive environment, and in the absence of erosion, limits the
corrosion rate. However, if stressors causing erosion are present, the layer of iron oxide will
dissolve and the uncorroded metal surface will again be exposed to the corrosive environment,
and piping corrosion will continue. Thus, the continuous process of oxide growth and
dissolution leads to thinning of the pipe wall and ultimately to a catastrophic failure when the
pipe is subject to a pressure pulse of large magnitude. Figure 18 presents a simple model
describing the phenomena occurring during erosion-corrosion (Sanchez-Caldera 1984).
The factors affecting the erosion-corrosion rate include the following:
• Piping configuration
• Feedwater temperature
• Bulk-flow velocity
• Turbulence
• pH level
• Oxygen content
• Impurities
• Piping material
Carbon steel components with less than 0.1 weight percent Cr are susceptible to erosion-
corrosion damage. Although erosion-corrosion is a greater concern in PWR feedwater piping,
steam generator components have also experienced damage from this mechanism.
Erosion-corrosion of the thermal sleeve at Diablo Canyon Unit 1 was recently reported
(USNRC 1993). The carbon steel J-tubes and feedrings within RSGs have also experienced
significant erosion-corrosion-induced wall thinning. The affected J-tubes have been repaired
or replaced with Alloy 600 J-tubes.
56
Erosion-corrosion damage has been reported to some of the carbon steel primary side
divider plates in the CANDU steam generators, as well as fatigue damage to the divider plate
bolts (also carbon steel). (The primary side divider plate is located below the tubesheet in the
lower plenum of the RSGs.) The erosion-corrosion of the plate and fatigue of the bolts caused
increased divider plate leakage and excessive bypass flow, which decreased somewhat the
performance of the steam generators. Fatigue of the bolts may also lead to loose parts damage
to the tubesheet.
4.6. WWER COLLECTOR, SHELL, AND FEEDWATER DISTRIBUTION SYSTEM
Although the WWER tubing has been relatively trouble free, stress corrosion cracking
of the WWER-1000 cold collectors and erosion-corrosion of the WWER feedwater
distribution systems has occurred The stress corrosion cracking of the collectors is discussed
first, followed by a brief discussion of the feedwater problems.
4.6.1. Stress Corrosion Cracking of the WWER-1000 Collectors
In contrast with the vertical tube bundles and horizontal flat tubesheets used in the West,
the WWER steam generator tube bundles are horizontal and are attached to the walls of two
vertical cylindrical collectors or headers. Primary coolant from the reactor core region enters
through the inlet (hot) collector, passes through the U-shaped tubing, and leaves through the
outlet (cold) collector. The WWER-1000 collectors are fabricated from low alloy steel and
clad on the inside with austenitic stainless steel. The hot and cold collectors are similar with
normal operating temperatures at 320°C and 290°C, respectively. The inner diameter of the
WWER-1000 collectors is 834 mm and the wall thickness is 171 mm.
Higher than normal radioactivity levels were observed in the secondary system of South
Ukraine Unit 1 in late 1986. It was determined that three adjoining ligaments in the cold
collector of one of the four steam generators had developed through-wall cracks resulting in
failure of the tube-to-collector inside surface cladding welds and significant leakage of
primary coolant into the secondary system (Titov 1991). This steam generator had been in
operation for less than one year. As of July 1993, 33 steam generators at eight WWER-1000
plants had been replaced because of failure or the potential of failure of the cold collectors
(Koryakin 1993, Titov 1991). These replacements occurred at only 3-25% of the design
lifetime (240 000 h).
Cracking and potential rupture of the WWER collectors is of concern not only because
of the economic losses associated with repairing or replacing these steam generators, but also
because of public safety. Radioactive primary coolant could be discharged to the environment
via the main steam atmosphere dump valves if they stick open. Worst case calculations
suggest that about 200 tonnes of primary system, steam generator, and emergency core
cooling water might be released. Also, long term cooling might be lost if the atmospheric
dump valves do not close properly because there are no isolation valves on the atmospheric
dump valve lines (IAEA 1994).
Metallographic examination of failed collector material "showed that the cracks were
corrosion-induced, mechanical in nature, initiating and propagating from the secondary circuit
57
side, at first via an intercrystalline and then via an intergranular mechanism" (Titov 1991).
The maximum crack length (as a sum of the lengths of the affected ligaments) on the
secondary side was about 1000 mm. The maximum through-wall crack length on the primary
side was about 10-15 mm. To date, cracks have been found only in the cold collectors.
However, "indications" have also been reported for the hot collector (IAEA 1994). Three
types of cracks have been observed: satellite cracks with widths up to 0.1 mm and lengths
up to 1 mm; planetary cracks between two adjacent holes with widths up to 0.5 mm, lengths
across the ligament, and depths up to 30 mm; and arterial cracks through several (up to 30)
holes with lengths up to 1000 mm, widths more than 0.5 mm, and depths through the wall
(171 mm). The maximum crack propagation rate was six ligaments within one operating cycle
(approximately 18 months).
The metallographic examinations also showed that the cracks usually started at a crevice
between the collector hole and a non-expanded tube, near the non-perforated zone (V-
configuration) of the collector. The cracks start at pits and grow across the ligaments first,
further growth occurs through the wall. The wall is penetrated only after cracking of several
ligaments. Ductile cladding failure occurs after the cracks penetrate the collector wall.
Investigation and analysis of the design, fabrication, operational loads, and water
chemistry conditions led to the following findings (Titov 1991, Titov 1993):
(a) The tubing in the steam generators with collector cracking had been explosively
expanded into the collectors using "rigid" charges. This procedure led to
deformation of the collectors, seizure of the upper part of the collector in the steam
generator vessel flange, and residual stresses near yield in the collector ligaments.
(b) The collector hole drilling techniques coupled with the explosive tube rolling led
to the formation of a layer of embrittled, highly cold work material on the inside
surface of the collector holes, which was sensitive to cracking.
(c) Crevices with depths up to 20 mm were present due to under-expansion of the

tubes. These crevices collected impurity deposits which promoted stress corrosion
cracking. The deposits in the cold collector crevices tended to be porous, whereas
the deposits in the hot collector crevices were generally dense enough to prevent
water ingress.
(d) The low alloy steel used for the WWER-1000 collectors undergoes strain ageing
at about 290°C. It is also more susceptible to stress corrosion cracking at
temperatures below 280°C than at higher temperatures.
(e) Abnormal secondary water chemistry conditions accelerated the cracking process,
especially a drop in pH to acid conditions (as low as 4.3), and significant periods
when the chlorine ions ranged from a few hundred to a few thousand ppb rather
than the specified less than 150 ppb (Gorbatykh 1993, Martynova and Mamet 1991,
Rassokhin et al. 1992). Excessive oxygen due to aerated auxiliary feedwater and
copper from the condenser tubes may also have contributed to the problem (IAEA
1993).
58
(f) The steel fabrication process may have created MnS inclusions which acted as sites
for crack initiation.
To improve the performance of steam generators already in operation, the following

changes were made: release of the collector upper racks, low temperature heat treatment
(450°Q of the collector perforated zone, and improvements in secondary water chemistry.
These changes helped but were not fully effective. For steam generators fabricated but not in
operation, a high temperature heat treatment at 650°C was conducted along with full depth
hydraulic expansion of the tubes. The new WWER-1000U design will probably use titanium
stabilized austenitic stainless steel in the perforated regions of the collectors rather then low
alloy steel and the tubes will be expanded hydraulically.
The WWER-1000 steam generators fabricated in the Czech Republic by Vitkovice, J.

S. C. for the Temelin plant incorporated several improvements to address these problems. All
their Type 10GN2MFA low alloy steel was doubly vacuum treated to minimize the gas
concentrations and secure a homogeneous chemical composition. The phosphorus and sulphur
contents were reduced. The collectors were forged so as to suppress macrosegregations on
their inner surfaces. The tubing was expanded to the collector wall by a hydraulic expansion
process which ninirnized the residual stresses and crevices.
4.6.2. Erosion-Corrosion of the Feedwater Distribution System
As discussed in Section 4.5.4 above, erosion-corrosion is a flow-assisted corrosion

mechanism where damage caused by erosion-corrosion is higher than damage attributed to
erosion or corrosion alone. The factors affecting the erosion-corrosion rate include piping
configuration, feedwater temperature, pH, bulk-flow velocity, turbulence, oxygen content,
impurities and material.
A current issue involves the erosion-corrosion of the WWER-440 and WWER-1000

feedwater distribution systems. The WWER-440 feedwater distribution system is shown in
Fig. 4 and consists of a feedwater pipe which enters the steam generator vessel on the side
opposite the hot collector in the steam region, travels across half the tube bundle, and then
travels down to about the center of the tube bundle where it connects via a tee joint with a
horizontal manifold The horizontal manifold is equipped with a number of nozzles directed
down, through which sub-cooled feedwater is injected into the corridor between the two sides
of the tube bundle to mix with the saturated liquid. The system was originally fabricated with
mild carbon steel. Flow-assisted corrosion of the nozzles has occurred at a number of plans
including Dukovany, Paks and Rovno. The damage has ranged from modest wall loss to
complete nozzle destruction (the nozzles closest to the tee tend to be more damaged). Erosion-
corrosion of the tee joint has also been observed, which could lead to cold feedwater spray
onto the hot collector.
Loss of the feedwater distribution nozzles is not considered a major safety issue because
experiments conducted at OKB Gidropress have shown that the feedwater flow distribution
is still adequate. However, the missing parts may cause fretting damage to the steam generator
tubes or damage the valves in the blowdown lines (only 2 of the 13 missing feedwater
distribution nozzles at Paks have been found). Also, continued erosion-corrosion of the system
will eventually destroy the tee joint.
59
In response to this problem, OKB Gidropress has designed a new WWER-440 feedwater
system which has similar geometry but is made of titanium-stabilized austenitic stainless steel.
This new system has been installed in the Rovno steam generators and will be installed in the
Paks steam generators. Another retrofit design prepared by Vitkovice in the Czech Republic
is characterized by a manifold above the water level and feedwater distribution through long
downconers into mixing boxes situated at the level of the previous feedwater manifold. This
design has been installed in the Dukovany steam generators (16 steam generators) and one
Bohunice steam generator. A slightly different upper feedwater system was installed in
another Bohunice steam generator. The tee joints were repaired at Loviisa in 1989-1990.
Later, on finding extensive damage to the feedwater distribution nozzles, a programme of
feedwater system piping replacement was begun at Loviisa.
Erosion-corrosion of the WWER-1000 steam generator feedwater distribution systems

may also be a problem, and alternate designs and materials are being evaluated by OKB
Gidropress.
4.6.3. Failure of Cbllector Cover Bolts
The WWER steam generator collectors are sealed at the top with covers (plates) which
are bolted to thin flanges around the top of the collectors (see Figs 4b and 5b). On 24 January
1982 all twenty bolts holding the cover on the hot collector in Steam Generator Number 5
at Rovno Unit-1 broke during a reactor power increase from 75% to 82%. The cover blew
off (lifted), creating a break area around the collector circumference with an equivalent
diameter of about 120 mm. The primary coolant system pressure dropped rapidly and the
reactor was automatically scrammed at 12 seconds. All three trains of emergency core cooling
started shortly thereafter. At 13 minutes, the operators shut off the reactor coolant pump on
Loop 5 and attempted to close the isolation valve but it would not fully close (the primary
coolant pressure was about 40 atmospheres).
Between 30 and 39 minutes the operators were able to improve the leak tightness of the
Loop 5 isolation valve, but also noted that Loop 3 was leaking. Eleven of the twenty bolts
on the hot collector cover in Steam Generator Number 3 were later found to be broken. The
operators shut down the Loop 3 reactor coolant pump and tried to close the Loop 3 isolation
valve. It initially closed only 50% of the way. The result of these actions (full isolation of
Loop 5 and partial isolation of Loop 3 and full emergency core cooling flow) was that the
primary system pressure increased to 105 atmospheres at 39 minutes and then all twenty bolts
on the cover of the hot collector in Steam Generator 1 broke. (Also, four of the twenty bolts
on the cover of the hot collector in Steam Generator 4 broke at some point.) The primary
coolant system pressure then dropped back about 40 atmospheres within about 1 minute. At
65 minutes, there were indications that some of the steam generators were overfilled and there
was water in the steam lines. Eventually, all four defective steam generators were isolated and
the plant was cooled using Loops 2 and 6 only. Altogether, about 1100 tonnes of primary
coolant and emergency core cooling water was lost to the secondary side and about 20 tonnes
were released to the environment along with about 17 curies of radioactive material
(Solovyev, 1992).
60
Inspection of the bolts after the accident determined that the failures probably occurred
as a result of corrosion-fatigue damage. The bolting material chemical and mechanical
properties were within specification, however, there was some non-uniformity in yield
strength (56-67 kg/m2) and hardness (19-27 Rockwell). Forty per cent of the fracture surfaces
had a clearly visible striated structure characteristic of fatigue damage. There were differences
in grain size ranging from 3-9 degrees BALL and carbide inclusions. The "character of the
fracture surfaces was brittle" with numerous inter- and trans-granular microcracks. The breaks
occurred in the transition region from the threaded to non-threaded material or in the first few
threads. Some of the microcracks appeared to have been present for a considerable period of
time.
Due to wear of the top cover seals, there had been primary to secondary coolant system
leakage from the hot collector covers in Steam Generators 1, 3, 4 and 5 and the bolts had
been screwed down very tight the previous year, creating high tensile stresses. Other possible
reasons for the bolt cracking include water level oscillations and splashing on the secondary
side which caused thermo cycling and fatigue damage and may have caused an accumulation
of impurities in the bolt region, a poor choice of bolt lubricant, and high chloride levels on
the secondary side (Solovyev 1992, IAEA 1995).
Corrective measures at Rovno included a new procedure for tightening the bolts, a
change in the stud lubricant from molybdenum sulphide to copper-graphite, a change in the
secondary side chloride limits from 500 ppm to 50 ppm, and better secondary side water level
control. Also, the bolts and covers on all the Rovno steam generator collectors were replaced.
Other WWER-440 plants have also implemented nitrogen-16 monitoring on the main steam
lines in response to this accident (IAEA 1995).
4.7. SUMMARY OF CURRENT WORLD EXPERIENCE
The status of the western (PWR and CANDU) steam generator tubing degradation is
summarized first using information from EPRI (1994 and 1995a). That is followed by a short
summary of the tube rupture experience and the risk impacts of tube ruptures. The status of
the WWER steam generators is then discussed.
Figure 19 shows the percentage of the total number of PWR and CANDU plants that
have plugged defective tubes in a given year. Data from 1975 to 1994 are plotted In recent
years, about one half of the PWR and CANDU nuclear power plants in the world were
plugging steam generator tubes in any given year. This implies that about one half of the
PWR and CANDU plants were operating with tubing defects near or beyond the national
limits in any given year. Figure 20 shows the steam generator tubes plugged per year as a
percentage of the total number of steam generator tubes in service, hi recent years, the
percentage of tubes plugged per year has been about 0.30-0.34 per cent (of a total steam
generator tube population which topped 3.4 million in 1994). The total number of steam
generator tubes plugged per year during the last few years has ranged from about 8000 to
10 000 tubes. In addition, more than 55 000 steam generator tubes had been sleeved as of
December, 1993 and about 30 000-40 000 tubes were sleeved in 1994 and 1995.
61
60-1
76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
Year
FIG. 19. Percentage ofPWR and CANDU Nuclear Power Plants Plugging Steam Generator
Tubes (EPRI 1995a). Copyright 1996 Electric Power Research Institute; reprinted with
permission.
0.45-1
76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
Year
FIG. 20. Percentage of PWR and CANDU Steam Generator Tubes Plugged (EPRI 1995a)
Copyright 1996 Electric Power Research Institute; reprinted with permission.
62
Although an average plugging rate of 0.25-0.3 per cent per year may seem acceptable,
over a 40 year steam generator life this amounts to about 10-12 per cent of the available
tubes plugged Also, not all steam generators are degrading equally. Table Xm lists some of
the plants which have plugged significant numbers of tubes in the last few years. The causes
of steam generator plugging on a world wide basis are shown in Fig. 13. As discussed above,
the relative impact of the various tube degradation mechanisms on overall steam generator
performance has dramatically changed over time, hi 1993, PWSCC (22%), ODSCC (41%)
and fretting (5%) accounted for about 68% of all the tubes plugged The diversity and
persistence of the damage mechanisms suggest that no one remedy will resolve all the
problems and effective remedies are not easily found. However, it should be noted that there
have been far fewer tubing failures in the replacement steam generators than in the original
TABLE Xm. PWR AND CANDU POWER PLANTS WITH SIGNIFICANT TUBING
DEGRADATION IN RECENT YEARS
Plant Year # Tubes Plugged Cause

7 1990 548 PWSCC, IGAa
arley - 2
1990 617 AVB Fretting
Palisades
1990 624 PWSCC, IGAa
ASCO- 1
1991 468 ODSCC/IGA
St. Lucie - 1
1991 1580 ODSCC/IGA
Trojan
1991 600 PWSCC, IGA"
forth Anna - 1
1991 760 ODSCC/IGA
Beaver Valley - 1
1991 331 PWSCC
Bugey - 5
1991 400 ODSCC/IGA
Angra - 1
1992 306 ODSCC/IGA
ANO- 1
1992 354 ODSCC/IGA
D.C. Cook - 1
1992 527 ODSCC/IGA
forth Anna - 2
1992 612 Erosion/corrosion3
Oconee - 1
1992 353 ODSCC/IGA
ASCO- 1
V.C. Summer
•forth Anna - 2
Byron - 1
1993 740 PWSCC, IGA"
Catamba - 1
1993 431 ODSCC/IGA
ASCO -2
1993 440 ODSCC/IGA
Bruce - A
Secondary side.
63
equipment. Therefore, one would expect that the numbers of degraded and plugged steam
generator tubes will, at some point, begin to decline as more replacement steam generators
come on-line. As of December, 1993, a total of 50 steam generators at 18 nuclear plants in
Belgium, France, Germany, Japan, Sweden, Switzerland, and the USA had been replaced.
As discussed in Section 4.4, the widespread tubing degradation that has occurred in some
PWR steam generators had led to spontaneous single tube rupture events which have occurred
at a rate of about one rupture every two years. In addition, incipient tube rupture events have
been occurring at the rate of about once a year in recent years. Steam generator tube ruptures
due to loose parts damage, PWSCC and ODSCC are likely to continue to occur for a number
of years.
During a tube rupture transient, the reactor operators are expected to (a) maintain the primary
coolant subcooled, (b) minimize the leakage from the reactor coolant system to the defective
steam generator secondary side, and (c) minimize the release of radioactive material from the
damaged steam generator. The success of the reactor operators has been mixed; some were
slow to understand what was occurring, slow to start reducing power, and slow to isolate the
defective steam generator. Others reduced power and isolated the defective steam generator
promptly. Some operators were slow to cool and depressurize the primary system, others took
prompt action. The result was that the defective steam generators were overfilled in a number
of cases and more radioactive material was released to the environment than necessary.
Nevertheless, in all cases the plants were properly cooled down and the radioactive material
releases were small and well below regulatory limits.
Simultaneous rupture of a number of steam generator tubes is very unlikely unless

induced by a design basis accident such as a main steam line break. There have been no main
steam line breaks or other design basis accidents which might cause multiple steam generator
tube ruptures. However, such accidents are possible. Sophisticated analysis of such transients
in PWRs indicates that effective operator intervention and actions to throttle the emergency
core cooling injection and actuate the residual heat removal system will result in a successful
recovery from a main steam line break with up to about 15 induced steam generator tube
ruptures. More than about 15 induced steam generator tube ruptures produces a system
response where the reactor coolant system subcooling cannot be recovered prior to exhaustion
of the normally available emergency core cooling water (which in some plants can be
replenished from outside sources). Some seven hours of emergency core cooling are available
for a main steam line break with one induced tube rupture, whereas only about two hours of
emergency core cooling are available for a main steam line break with 15 induced tube
ruptures. (These are typical numbers for a Westinghouse-type three loop plant and assume
optimum throttling of the emergency core cooling injection.) Clearly, the reactor operator
actions must be prompt and effective: this was not always the case during the previous single
spontaneous steam generator tube ruptures. If the reactor operators do not properly throttle
the emergency core cooling injection or replenish the storage tank during a main steam line
break accident with 15 induced steam generator tube ruptures, the emergency core cooling
water supply will be exhausted within an hour and the core will be uncovered and start to
melt in about seven hours.
64
To summarize, there are a number of PWR and/or CANDU steam generators operating
with tube defects at or near the national limits at any given time. Spontaneous single tube
ruptures have and will probably continue to occasionally occur. The radioactive material
releases associated with these events have and will continue to be small and well below
regulatory limits. A design basis accident such as a main steam line break might induce some
multiple tube ruptures. It is likely, but not certain, that the reactor operators can successfully
cope with the transient.
hi contrast with some of the PWR and CANDU steam generator tubing, the WWER
tubing has been relatively trouble free. However, the collectors in the WWER-1000 steam
generators have been a problem. As of July 1993, 33 steam generators at eight WWER-1000
plants had been replaced because of failure or the potential of failure of the cold collectors.
Unfortunately, many (most) of the replacement steam generators are not significantly different
from the original equipment so additional collector cracking is expected. The collector cracks
can be large and the crack propagation rates high. Cracks up to 1000 mm in length have been
found and crack propagation rates up to six ligaments per operating cycle have been observed.
65
OPERATIONAL GUIDELINES
This section describes a set of operational guidelines which will help nuclear power plant
operators prevent, or at least minimize, steam generator ageing degradation and thereby
maximize component life. Implementation of these measures is expected to be considerably
less costly than repairing or replacing steam generators and may provide the additional
assurance required to operate some steam generators for additional time.
The operational procedures are grouped into the following topic areas:
- Primary coolant system water chemistry control parameters

- Secondary coolant system water chemistry control parameters
- Measures to control secondary-side impurity incursions
- Measures to remove secondary-side impurities
- Measures to control steam generator deposits
Although this section provides general guidance, detailed operating procedures must be
developed to suit the specific conditions and requirements of specific plants. Given the variety
of materials and environmental conditions encountered in practice, the detailed operating
procedures will vary somewhat from plant to plant.
A very important aspect of steam generator ageing management is the use of a

comprehensive inspection and monitoring programme and appropriate fitness-for-service
guidelines to assess the current and future safety state of these components. These topics are
discussed in Sections 6 and 7 of this report.
5.1. PRIMARY COOLANT SYSTEM WATER CHEMISTRY CONTROL

PARAMETERS
The purpose of the PWR primary coolant system water chemistiy programme is to
protect the fuel rod cladding from excessive oxidation and crud buildup and provide reactivity
control for the reactor. PWR primary coolant system water chemistry which meets these
objectives has no effect on steam generator degradation. Nevertheless, the PWR primary
coolant water chemistry control parameters are discussed in this section for completeness.
PWR Steam Generators. The important parameters of the PWR primary reactor coolant
chemistry are the boric acid, lithium hydroxide, and hydrogen concentrations, and the
resulting pH level. A minimum high temperature (~300°C) pH of 6.9 (pH300 = 6.9) is required
to avoid heavy crud deposits on fuel rods, which can cause accelerated corrosion of fuel rod
cladding and increased radiation fields [Lott et al. 1992]. Some test results show that
operation at a pH300 of 7.4 results in less crud deposits than that at 6.9. For current PWR
operation, the typical range of pH300 is 6.9 to 7.4. The pH300 for most of the Elecrricite de
France (EDF) plants is 6.9, and the pH is fixed at 7.2 if the cycle duration is 18 months.
66
Boron is added in the form of boric acid (H3BO3) as a neutron absorber for reactivity
control. The boric acid concentration is changed throughout a reactor cycle to compensate for
other changes in reactivity and is not varied independently. The boron levels are relatively
high (1000 to 2000 ppm) at the beginning of the fuel cycle. Then, they are gradually reduced
by 100 ppm/month. The concentration of lithium hydroxide (LiOH) is co-ordinated with the
boric acid concentration to achieve the desired pH of approximately 6.9 or higher at operating
temperature. At the beginning of the fuel cycle, the typical lithium level is about 4 ppm for
a boron level of 2000 ppm, and then it is reduced as the boron level reduces [Lott et al.
1992].
Hydrogen is added to the primary coolant to suppress the buildup of oxygen from
radiolysis. A hydrogen concentration of 25-50 crrrVkg has typically been used. Recent EPRI
sponsored studies indicate that increasing the hydrogen concentration in the primary coolant
increases the rate of primary water stress corrosion cracking (PWSCQ. Consequently, EPRI
is encouraging utilities to maintain hydrogen concentrations near the low end of the specified
range (i.e., 25- 35 cm3/kg) [German 1989].
The Revision 2 EPRI guidelines for PWR primary coolant system water chemistry are
listed in Table XTV (EPRI 1990). EPRI is also about to issue a third revision of their
TABLE XTV. EPRI PRIMARY COOLANT SYSTEM WATER CHEMISTRY

GUIDELINES FOR POWER OPERATION (REACTOR CRITICAL)
Control Sample Typical Value Action Level

Parameter Frequency
1 2 3
Chloride, ppb 3/wk« <50 — >150 >1500
Fluoride, ppb 3/wk^ <50 — >150 >1500
Lithium, ppm 3/wk^ Consistent with — — —
Station Lithium
Programme
Hydrogen, 3/wk<c) 25-50<« <25 <15 <5
cc(STP)/kg H2O >50
Dissolved 3/wk<a) <5 — >100 >1000
Oxygen, ppb
These frequencies are a nindmumbased on Standard Technical Specifications. Typical industry frequencies are daily.
An increased frequency of sampling is recommended during operations that may significantly impact the lithium
concentration (i.e., feed and bleed).
An increased frequency of sampling is recommended during operations that may significantly impact the hydrogen
concentration (i.e., feed and bleed, purging of pressurizer vapor, etc.)
Maintain near the low end of this range.
67
guidelines which will include the following changes: sulphate is added as a control parameter
(50 ppb); the Action Level 1 for chlorides and fluorides will be 50 ppb, each; and the
limitation on hydrogen control at 25-50 cc/kg will be removed for plants with steam
generators susceptible to PWSCC. Also, there are some minor changes to the pH optimization
principles, and the Level 1 definition. (The Level 1 value is now the value outside of which
data or engineering judgement indicates that long-term system reliability may be affected,
thereby warranting an improvement of operating practices.)
CANDU Steam Generators. The pH in the CANDU primary system is controlled by

addition of lithium hydroxide or by using lithiated nixed bed ion columns. Boron and boric
acid are not used in the CANDU primary system because the on-line refueling feature
eliminates the requirement for a borated fuel shim.
WWER Steam Generators. The WWER primary reactor coolant chemistry is a reducing,
weak alkaline chemistry treated with the addition of ammonia, potassium, and boric acid. The
allowable at power pH and dissolved hydrogen, oxygen, ammonia, chloride, fluoride, iron,
oil, copper, and boric acid concentrations are listed in Table XV for the WWER-440 plants
with corrosion resistant stainless steel cladding on the inside surface of the reactor pressure
vessel, the WWER-440 plants without reactor pressure vessel cladding, and the WWER-1000
plants. Valves for the Temelin plant in the Czech Republic are also listed in Table XV.
TABLE XV. WWER PRIMARY WATER CHEMISTRY
Control parameters WWER-440 WWER-440 Temelin WWER-1000

with stainless steel clad without stainless steel
reactor vessel cladding on the
reactor vessel
pH at 25°C >60 60-102 57-102" 59-103
Ammonia, ppm >50 >50 >50" >50
Hydroeen 27-54 27-54 3-6 27-54
(at 0°C, 0 1 MPa), ppm
Dissolved oxygen, ppm <0 01 <0005 <001 <0005
Chloride and fluonde, ppm <0 1 <01
Corrosion products in terms of iron at steady <0 2 <02 <02
state operation, ppm
Oil, ppm <005 <005
Copper, ppm <002 <002
Bone acid, depending on core reactivity margin, 0-8 0-90 08 0-100
g/kg
Total iodine isotopes radioactivity at the time of <3 7X108 <37X108
sampling, Bq/1
The lithium to boron ratio is co-ordinated to maintain a pH of 6 9 If the lithium decreases to 2 2 ppm, it is held at 2 2 ppm until a pH of
7 4 is reached
b
Normallvat 10
68
5.2. SECONDARY COOLANT SYSTEM WATER CHEMISTRY CONTROL
PARAMETERS
A secondary coolant system water chemistry programme should be established that limits
the steam generator water impurity concentrations to certain specified values. The programme
should identify all the required continuous and grab samples, specify the accuracy and
frequency of the measurements, and specify the chemistry levels that initiate various
corrective responses up to and including plant shutdown.
Recirculaing Steam Generators. The PWR secondary water chemistry was based on co-
ordinated phosphate additions to provide a buffering system until about 1974 in the USA,
1975 in Japan, and the mid 1980s in Germany. However, phosphate chemistry results in
concentrations of chemicals in a sludge pile, which in turn causes general corrosion of the
outside tubing surfaces. After several laboratory studies and field observations, almost all
PWRs have now been switched from phosphate chemistry to an all-volatile treatment to
mitigate the steam generator tube wastage problem. (However, one plant, Doel Unit 4, started
operation with an all volatile treatment and switched to phosphate and one Spanish plant, Jose
Cabrera, started with phosphate in 1968 and one Argentine plant, Atucha-1, started with
phosphate in 1974; both remain on phosphate.)
The initial all-volatile treatments used a low molecular weight amine, ammonium
hydroxide (NHjOH) (ammonia) along with hydrazine (NjFL,) as a scavenger of trace quantities
of dissolved oxygen. Recently, there has been a trend away from using ammonium hydroxide
toward the use of morpholine (C4H8ONH), which is somewhat less volatile, in conjunction
with hydrazine. This results in even higher pHs in many parts of the secondary system and,
therefore, further reduces erosion-corrosion in the feed train and wet steam piping. This, in
turn, reduces sludge buildup in the steam generators. By the end of 1990, about 30% to 40%
of the US PWRs and 75% of the French PWRs were converted to a morpholine all-volatile
treatment.
A few US plants have converted to a combination of boric acid and morpholine. About
30% of the Japanese PWRs are converted to boric acid without morpholine. Some plants have
also tried ethanolamine [NH2(CH2)2ON]. Laboratory studies indicate that the addition of boric
acid will prevent denting and IGSCC/IGA initiation in alkaline environments that would
otherwise cause damage [Wood 1990]. Further findings indicate that adding boric acid after
crack initiation in alkaline environments reduces the rate of crack propagation by a factor of
8 to 10. Boric acid can be added during normal plant operation and also during tubesheet
crevice flushing operations performed during shutdown. The process has been fully qualified
for compatibility with steam generator components by Westinghouse and ABB-Combustion
Engineering. By the end of 1990, 32 power plants had accumulated a total of about 60 years
of operating time with boric acid
hi addition, some plants in the USA are operating with elevated hydrazine concentrations
(greater than 100 ppb) which increases the resistance of the steam generator tubing to
IGA/IGSCC and pitting. Hydrazine is a reducing agent that decreases the electrochemical
corrosion potential when oxidants are present and it is now recognized that high hydrazine
concentrations provide better assurance that reducing conditions are being maintained.
69
TABLE XVI. WATER CHEMISTRY CONTROL PARAMETERS FOR RSGS AT >5% POWER (EPRl 1993C)
Parameter Measurement frequency Median value Action level l ( b ) Sample location

pH @ 25°C (station with no Continuous 9.3-9.8 (a) Slowdown
copper alloys) (a) Feedwater
pH @ 25°C (station with Continuous 8.8-9.2 >9.2 Condensate
copper alloys) (a) Slowdown
(a) Feedwater
pH agent Daily — (a) Feedwater
Dissolved O2, ppb Continuous <3 >5 Feedwater
>10 Condensate
Hydrazine, ppb Daily (a) <100 Feedwater
Total iron, ppb Weekly — >5 Feedwater
Total copper, ppb Weekly — >1 Feedwater
Cation conductivity pS/cm Continuous 0.15 >0.8 Slowdown
(ammonia AT) <0.2 Feedwater
Cation conductivity faS/cm Continuous 0.5 >0.8 Blowdown
(non-ammonia AT) <0.2 Feedwater
Sodium, ppb Continuous ~2 >20 Blowdown
Chloride, ppb Daily ~2 >20 Blowdown
Sulphide, ppb Daily ~3 >20 Blowdown
"Per station ^11

rci 3iaui7ii pH programme
[jiugimiiiii&.
b
Corrective actions should be implemented as soon as possible. If the parameter is not below this value in one week, go to Action level 2.
TABLE XVH. SIEMENS/KWU GUDDI FOR SECONDARY WATER CHEMISTRY
Control Parameter Diagnostic

Parameter
Feedwater
PH >9.8 -
Cation Conductivity (25°Q uS/cm <0.2 -
Oxygen ppb <5 -
Hydrazine ppb - >20
Specific Conductivity (25°Q uS/cm - >15
Steam Generator Slowdown
pH - >9.5
Cation Conductivity (25°Q uS/cm <1 -
Sodium ppb <50 -
Main Condensate
Cation Conductivity (25°Q uS/cm - <0.2
Oxygen ppb - <20
HotweUs
Sodium ppb - < 1
(Difference to Reheated Steam)
Reheated Steam
Make-up Water
Specific Conductivity (25°Q uS/cm - <1
Chloride ppb - <50
Silica (in form of SiO2) ppb - <20
However, the thermal decomposition of some of the hydrazine to ammonia can cause
problems in plants with copper alloy material in the secondary coolant system. The ammonia
causes accelerated corrosion of the copper alloy and may adversely impact dernmeralizer
operation. Nearly all of the Siemens/KWU plants are operating with high hydrazine
concentrations (greater than 20 ppb specified and typical values of 80-200 ppb) and high pH
values (greater than 9.8 specified and typical values of 10.0 to 10.2). The copper alloy
condenser tubing in the older Siemens/KWU plants was replaced with either stainless steel
or titanium tubing when the plants were converted to high hydrazine, high pH water
chemistry.
71
Because an all-volatile treatment does not buffer the system, the water chemistry needs
to be constantly monitored and corrected. Water chemistry control guidelines developed by
the EPRI Steam Generator Owners' Group require continuous (or daily) monitoring of cation
conductivity, chloride, sodium, sulphate, pH, ammonia, dissolved oxygen, hydrazine, copper,
and iron [Mundis 1983, EPRI 1982, EPRI 1993c]. The Steam Generator Owners' Group
guidelines also establish very low levels of acceptability for impurities because the water
chemistry at an all-volatile treatment plant is more sensitive to small quantities of impurities
than the water chemistry at a plant using phosphate. In-leakage of seawater or brackish water
due to condenser tubing defects, impurities in the feedwater, impurities released from the
condensate polishers, and resins released by the condensate polishers can all cause upset
conditions. The control parameters for greater than 5% power developed by the EPRI Steam
Generator Owners' Group are listed in Table XVI. Note that the pH at 25°C in plants with
no copper alloys in the secondary system should be higher (9.3 to 9.8) than the pH in plants
with, for example, copper alloy condenser tubes. These control parameters are to be used as
part of an approach where each plant develops a water chemistry optimization programme
based on corrosion history, cooling water chemistry, secondary system design, and operating
trends. The control parameters for the secondary water chemistry in Siemens/KWU plants are
somewhat different from the EPRI guidelines and are listed in Table XVEL EDF also uses
TABLE XVm. EDF GUIDELINES FOR FEEDWATER CHEMISTRY", POWER >25%,

AMMONIA (AVT)
Parameter Unit Typical Value Limiting Value Frequency of Remarks

Measurement
pH @ 25° 9.0 - 9.2 Continuous WithCu

9.6 - 9.8 Continuous Without Cu
Conductivity uS/cm 2.7 - 4.2 Occasional With Cu

@ 25°
10- 17 Without Cu
Ammonia ing/kg in NH4 0.25 - 0.5 Required quantity S WithCu
2-5 Without Cu
Oxygen Hg/kg <5 Continuous
Hydrazine ig/kg >5 Continuous With Cu
50 5 - 100 Without Cu
Suspended Iron ^g/kg < 10 Quarterly

Campaigns
Suspended Hg/kg <5 Quarterly

Copper Campaigns
Soluble Copper ug/kg <5 Quarterly

Campaigns
''Taken from EDF Instruction 83.032 Ind.l - TE/M 1628 Ind E "Regies generates d'exploitation".
72
TABLE XIX. EDF GUIDELINES FOR FEEDWATER CHEMISTRY", POWER >25%,
AVT: MORPHOLINE
Parameter Unit Typical Value Limiting Value Frequency of Remarks

Measurement
pH @ 25° 9.1 - 9.3 Continuous With Cu
9.1 - 9.7 Continuous Without Cu
Conductivity uS/cm 3-5 Occasional WithCu

@ 25°
3 - 13 Without Cu
Morpholine mg/kg 4- 8 >4 S With Cu
6-8 >4 Without Cu
Ammonia mg/kg in NH4 <0.3 <0.5 S With Cu
<3 Without Cu
Oxygen Mg/kg <5 Continuous
tfydrazine Mg/kg >5 Continuous WithCu
50 5- 100 Without Cu
Suspended Iron ig/kg < 10 Quarterly

Campaigns
Suspended -ig/kg <5 Quarterly

Copper Campaigns
Soluble Copper ig/kg <5 Quarterly
Campaigns
Taken from EDF Instruction 83.032 Ind.l - TE/M 1528 Ind E "Regies generates d'exploitation".
somewhat different guidelines for the primary and secondary water chemistry than
recommended by EPRI. The EDF values for the secondary side feedwater are listed in Tables
XVin and XIX; for plants at powers greater than 25% and ammonia or morpholine water
chemistry, respectively.
Some plants are also trying to balance the ratio of cations and anions in the crevices,
thereby preventing the formation of highly alkaline or acidic conditions. This technique is
known as molar ratio control. The practice of molar ratio control is based on the assumption
that the crevice pH can be modified by controlling the ratio of strong acid anions and strong
base cations in the steam generator bulk water. This, of course, requires an understanding of
the relationship between the bulk water chemistry and the crevice chemistry and estimates of
the hideout/hideout return fractions. The current version of the FJPRI PWR secondary water
chemistry guidelines (FJ'RI 1993c) suggests that plant operators consider implementing molar
ratio control at plants where ODSCC is a serious concern. This recommendation is based on
the fact that the available laboratory data indicate that IGA and IGSCC crack rates are lowest
at or near a neutral pH (5 to 9). However, there can be problems associated with molar ratio
73
TABLE XX. CHEMISTRY CONTROL PARAMETERS FOR ONCE-THROUGH STEAM GENERATORS
AT POWER
Initiate
Parameter Frequency Normal value action Action response
pH (a)
Hydrazine, ppb Daily >3x[02]b >3x[OJ Return to normal value within
(20 ppb mm) <20ppb 24 hrs. or be in hot standby
within an additional 24 hrs.
Dissolved oxygen, ppb Continuous <3 >3 Return to normal value within
1 week or reduce power to
-30% until source is identified
and isolated.
Sodium, ppb Continuous <3 >3 Return to normal value within
100 hrs or be in hot standby
Return to normal value within
>6 24 hrs or be in hot standby
Be in hot standby within
24 hrs.
Chlonde, ppb Daily <5 >5 Return to normal value within

24 hrs. or be in hot standby
within 24 hrs.
>20 24 hrs.
"Corrected" cation Continuous <D.2 X).2 Return to normal value within
conductivity, uS/cm 100 hrs or be in hot standby
within an additional 24 hrs
>0.5 24 hrs or be in hot standby
24 hrs.
Immediately go to hot
standby.
>2.0
Silica, ppb Weekly Return to normal value within

>20 24 hrs or be in hot standby
Immediately go to hot
standby.
>50
74
TABLE XX. (COIL)
Initiate
Parameter Frequency Normal value action Action response
Total iron, ppb Weekly <5 >5 Should be returned to normal

value within 24 hrs.
Copped, ppb Weekly <1 >1 Should be returned to normal
value within 24 hrs.
Sulphate, ppb Daily <3 >3 Return to normal value within
within 24 hrs.
a
Per station pH programme.
b
Oxygen measured at condensate pump discharge.
c
May be deleted for all-ferrous systems.
control. First, lead assisted transgranular stress corrosion cracking reaches a maximum at near
neutral conditions. Second, if excessive amounts of chloride are added to the secondary
coolant, denting and pitting can be a problem in certain steam generators. Therefore, an upper
chloride limit of 5 ppb (10 ppb during transients) is recommended and molar ratio control
should only be used in steam generators with low lead levels.
Once-Through Steam Generators. All once-through steam generators use AVT water
chemistry. Phosphate chemistry has never been used in the once-through steam generators,
so phosphate wastage has never been a problem. However, the once-through design is
susceptible to sludge buildup around the lower support plate flow holes that restrict the
feedwater flow. Such flow restriction has forced some Babcock & Wilcox plants to derate by
as much as 30% at times. The once-through steam generator water chemistry control
guidelines developed by the Owners' Group are listed in Table XX.
CANDUSteam Generators. CANDU Steam Generator chemistry control is generally all

volatile: morphoUne/hydrazine (NjHJ for plants with copper alloys in the feedtrain, NH3/N2H4
for all ferrous plants. One CANDU station uses only morpholine and no hydrazine, another
station uses a combination of phosphate and morpholine/NjHt- The latter is also the only
CANDU plant with a full-flow condensate polisher. Most of the CANDU units operate with
a deaerator to reduce dissolved oxygen. One CANDU station has started the use of boric acid
to mitigate corrosion of the carbon steel tube supports, in particular, the U-bend supports.
Corrosion of these supports has contributed to the "growth" of the entire structure, creating
high local stresses leading to extensive stress corrosion cracking in the U-bend region. The
initial operating experience with boric acid has been good and the impact on other chemical
parameters small.
75
TABLE XXI. WWER SECONDARY WATER CHEMISTRY
Russian Plants:
Control parameters Feedwater below the high Slowdown

pressure preheater water
pH at 25°C 90 + 02 80-92
Specific cation conductivity. uS/cm <03', OS*5 <5C^' 590s
Sodium, ppb - 53004-' <500-<
Chloride, ppb - <150"
Iron, ppb
Copper1, ppb <5 <3'

2
Dissolved oxygen ppb
Oil, ppb <100
Hydrazine, ppb
Notes 1 Measured down the last low pressure preheater

2 Down the deaerator
3 If low pressure preheaters have stainless steel tubes
4 Novovoronezh Units 3 and 4
5 Kola Units 1, 2, 3, and 4
6 WWER-1000 plants
7 Prior to 1990 this was 500 to 1000 ppb
Czech Plants:
Control parameters Feedwater Slowdown Main
water condensate
pH at 25°C 97-10
Specific canon conductivity, uS/cm <012
Sodium, ppb
Chloride, ppb
Dissolved oxygen, ppb <5
Hydrazine, ppb 30-200
Diagnostic parameters
pHat25°C 94-97 10
Specific conductivity, uS/cm 12-25 25
Iron, ppb
Copper, ppb <2
Ammonia, ppm 35-12
Silicon, ppb «300
Fluonne, ppb
NH,, ppm 1446
SCV. ppb <4
Notes a. First action level X) 7, second action level >2, third action level >7
b First action level >20, second action level >100. third action level >500
c First action level X) 2, second action level >1 0
76
WWER Steam Generators. Ammonia and/or hydrazine water chemistry is used on the
secondary side of the WWER steam generators. The WWER secondary water chemistry is
designed to minimize (a) sludge buildup on the heat transfer surfaces, (b) corrosion and
erosion damage, and (c) waste water. The WWER control parameters and allowable values
for operation at any power level are listed in Table XXI. All parameters are measured at
standard conditions of approximately 25°C and 0.1 MPa. Ingress of ion exchanger resin and
products of the resin deterioration is not allowed Deviation from the control parameters is
addressed through the concept of action levels (three action levels are used).
5.3. MEASURES TO CONTROL SECONDARY-SIDE CHEMICAL IMPURITY

INCURSIONS
Various organic acid and ionic impurities promote corrosive processes such as
IGSCC/IGA, pitting, denting, and cold leg thinning in steam generator tubing. Impurities can
be minimized by preventing in-leakage of raw water from condenser tubes, appropriate
filtration of the condensate and feedwater, reducing the volume of make-up water, and
elimination of copper alloy parts in the secondary coolant system.
Condenser Integrity. In-leakage of raw water through defects in the condenser tubes is
an important cause of faulted chemistry conditions for RSGs which result in ODSCC, denting
of tubes in generators with carbon steel support plates and pitting at the top or within the cold
leg sludge pile region. To reduce condenser leakage, several plants have replaced their
admiralty brass condenser tubing with either titanium (seawater and freshwater sites), or
stainless steel tubing (freshwater sites only) with tube-to-tubesheet welds. Use of titanium
condenser tubing is a standard feature in the more recent plants; however, a number of
titanium tube leaks in condensers have occurred. The possible degradation mechanisms for
titanium tubing are high-cycle mechanical fatigue caused by flow-induced vibrations, damage
from loose parts such as broken turbine blade pieces, and hydriding. Additional design
modifications may be needed to reduce the flow-induced vibrations. Also, the high
temperature steam connections to the condenser should be bellows-type expansion joints to
prevent air in-leakage. For existing installations without titanium or stainless steel tubing and
welded joints, aggressive leak detection and correction methods should be used
Condensate Polishing System. A condensate polishing system can potentially be an

important part of an effective secondary side water chemistry programme. Condensate
polishers are a cost effective means of achieving operating water chemistry specifications
during startup. Use during operation can essentially eliminate the transport of corrosion
products into the steam generator. And, full-flow condensate polishing protects against major
chemistry excursions due to, for example, condenser tube leaks (Dooley et al. 1995).
However, condensate polishers "elute low concentrations of impurities that are known
or suspected corrodents for secondary cycle materials. Those impurities include sulfate,
sodium, chloride, soluble sulphonated and chlorinated organics, and resin fines that can
degrade to sulphates and organic species in the steam generator" (Dooley et al. 1995). Also,
accidental ingress of resins from these systems has caused aggressive chemical environments
in some plants. If such an event occurs, immediate corrective action is needed to remove the
aggressive chemicals from the secondary coolant. This may include shutting down the reactor,
77
followed by flushing the system numerous times to prevent damaging concentrations of
soluble products in the crevices. Therefore, there is disagreement over whether condensate
polishers should be installed, and if they are installed, when they should be operated. Some
plants are not equipped with condensate polishers, some plants polish only a small function
of the feedwater flow, some plants only operate their condensate polishers during startup,
shutdown or periods of condenser in leakage, and some plants polish all the feedwater all the
time.
Recycle of Slowdown Water. Careful make-up water chemistry control is also needed to
control the chloride content in the secondary water (in addition to controlling the in-leakage
of raw water through the condenser). Organic impurities in the make-up water decompose at
steam generator temperatures and produce additional chlorides (decomposition products).
More chlorides may be introduced into the steam generator than indicated by an ordinary
chemical analysis of the make-up water, unless the water sample is subjected to high
temperatures and pressures before the analysis. One of the approaches to reducing the chloride
and impurity input through the make-up water is to reduce the quantity of make-up water
used. A blowdown recovery system will purify and recycle blowdown water that is cleaner
in terms of chlorides and organic impurities than the usual supply of make-up water.
Control of Lead Contamination. Lead has been implicated in the accelerated stress corrosion
cracking of Alloy 600 steam generator tubing at a number of utilities. Of particular interest
are two of the oldest French stations at Fessenheim and Bugey where cracking has been
related to possible chronic lead problems, rather than lead source contamination as in Bruce
Unit 2 and Doel Unit 4. Lead is typically found at comparatively low levels in steam
generator deposits having been transported there from the secondary system.
Laboratory studies have also revealed the presence of lead accelerated stress corrosion
cracking of not only Alloy 600, but also for Alloy 800 and Monel 400 (Takamatsu 1994,
Rocher et al. 1994, Briceno et al. 1994, Miglin 1991 and Palumbo and King 1992). All the
alloys tested corroded in aqueous lead contaminated environments, however, the Alloy 600
material was most susceptible. The most severe environment in the French studies (Rocher
et al. 1994) was a lead contaminated caustic solution. However, the Spanish studies show that
these materials have a higher susceptibility for cracking in a lead contaminated AVT
environment (Briceno et al. 1994).
At present there is no quantitative correlation between the feedwater or blowdown lead

concentrations and the cracking susceptibility of steam generator tubes under a given set of
generating conditions (e.g. constant stress and other water chemistry parameters). Therefore,
the lead contamination levels in operating steam generators should be kept as low as possible
until more information is available and some "safe" level is defined. Effective operating
procedures should be used to prevent and control both acute and chronic sources of lead
contamination.
Removal of the Copper from the Secondary Coolant System. Copper oxide may act as
an oxidizing agent in pitting corrosion of tubes and corrosion of tube support plates resulting
in tube denting (by enhancing the formation of magnetite). Copper is transported to the steam
generators due to the corrosion of components with copper-bearing alloys in the balance of
78
the plant. Therefore, copper-bearing alloys should be removed from the secondary coolant
system and replaced with carbon or stainless steel components. Also, copper oxide can be
reduced to metallic copper by flushing in a highly concentrated hydrazine environment.
Controlling the dissolved oxygen and eliminating the ingress of air are also important in
mitigating the pitting and denting mechanisms.
Summary. Constant monitoring of the water chemistry and immediate corrective actions
are important in maintaining the quality of the secondary water. Plant modifications that
ensure the quality of the secondary water contribute to mitigating all the corrosion related
degradation mechanisms on the secondary side.
5.4. MEASURES TO REMOVE SECONDARY-SIDE IMPURITIES
The objectives of secondary-side cleaning are to remove the sludge and the various
chemical impurities and corrosion products located in and under the sludge, and to remove
the chemicals concentrated in the tube/tubesheet crevices as well as in the tube/tube support
annuli. These procedures help prevent and mitigate IGSCC, IGA, and pitting.
Steam generator flushes. As discussed in Section 4, impurities can concentrate in the

tubesheet and tube-to-tube support plate crevices, the sludge pile, and under freespan crud
(bridging) deposits. These concentrated chemicals have caused extensive ODSCC pitting, and
other damage in certain plants. One procedure that may help control this problem is crevice
flushing, which is accomplished by heating the fluid in the crevices, then depressurizing the
steam generator secondary side so that boiling occurs in the crevices. The process is repeated
a number of times (more than four). Steam bubbles formed within the crevices (or in porous
crud) tend to expel the impurity-laden water, thereby cleaning the crevice. The crevices can
be heated (especially on the hot-leg side) by running the primary coolant pumps. The
mechanical energy of the pumps is translated into an increase in the primary coolant
temperature which heats the crevices. Electric heaters may be necessary to get enough heat
into the crevices on the cold leg side. Steam generator flushing has been tried at a number
of temperatures, the most common of which is 150°C. Sludge lancng should be performed
before crevice flushing so that impurities are not washed down into the crevices during the
flushes. Flushing steam generator crevices by depressurizing the secondary side appears to
remove most of the sodium, some of the sulphate, and not much of the chloride. Also,
depressurization without a nitrogen overpressure is about as effective as using a nitrogen
overpressure (Dooley et al. 1995).
Steam generator soaks. Although it has been suggested that removing the impurity source
from the feedwater and then running the plant at full power for a time or reducing the powero
periodically without shutdown might flush some of the impurities from the crevices, neither
has been effective. In fact, hot soaks at temperatures of 90°C to 150°C without heat flux have
been found to be the most effective simple means of promoting impurity releases (hideout
return) from the crevices. Both laboratory experiments and field experience indicate that the
amount of trapped impurity release by soaking is comparable to removal by depressurization,
although a longer time is required for soaking. However, removal of aggressive chemicals
from the tubesheet crevices is hindered by sludge on the tubesheet. Therefore, hot soaking
should be carried out after sludge cleaning (Dooley et al. 1995).
79
5.5. MEASURES TO CONTROL STEAM GENERATOR DEPOSITS
Deposits on the tubesheet (sludge), tube supports and on the tube surfaces create crevices
where impurities may concentrate. Concentration factors of greater than 105 over bulk water
have been found in laboratory tests (Gonzalez and Spekkens, 1986). The source of secondary
side deposits is corrosion of the secondary side components. Factors which influence this
corrosion are elevated oxygen levels due to air in-leakage or poor deaerator performance, pH,
poor quality make-up water and, lay-up conditions. Another major source of deposits comes
from condenser in-leakage (discussed in Section 5.3). A major effort should be in place
during unit operation to try to prevent the generation, transport and subsequent accumulation
of deposits in the steam generators. Preventive measures include: minimizing air in-leakage,
operating at high pH, and removing corrosion products with a good blowdown system and
by use of condensate polishers (see Section 5.3). Corrective measures include sludge lancing
and chemical cleaning.
Reduce twbine/condenser/steam air in-leakage. Plant procedures should be in place to

reduce the leakage of air into the secondary system, which require the use of state of the art
detection and repair technologies. This will minimize ingress of oxygen and thereby minimize
corrosion of the secondary side components. Oxygen is an oxidant itself and also increases
the transport of other oxidants (ionic copper species and iron oxides such as hematite) which
participate in corrosion reactions in RSGs.
Steam Generator Lay-up. At temperatures below about 200°F, the steam generators
should be filled with de-oxygenated (<100 ppb 02), chemically treated water to minimize
corrosion. An amine should be used to keep the pH above 9.8 and a hydrozine concentration
above 75 ppm should be used to maintain a protective oxide film and a reducing environment.
Hydrazine is an oxygen scavenger and inhibits general and localized corrosion of ferrous
materials and reduces the pitting susceptibility of Alloy 600. The sodium, chloride and
sulphate concentrations should be below 1000 ppb during wet lay-up and below 100 ppb prior
to heatup. A positive nitrogen overpressure should be maintained during filling, draining, and
cold shut down to minimize oxygen ingress. During periods when the steam generator must
be drained for maintenance, nitrogen should be used to prevent contact between the steam
generator water and oxygen. The steam generator water should be nixed and sampled three
times per week until the parameters are stable, and weekly thereafter. Corrective actions
should include feed and bleed operations or draining and refilling the steam generator with
deoxygenated make-up water of the proper purity (EPRI 1993c).
During heatup (reactor coolant system temperature above 100°C and reactor power below
5%), the dissolved oxygen in the feedwater should be as low as possible and below 100 ppb
before exceeding 5% power. The hydrazine in the feedwater should be greater than three
times the oxygen concentration and 100 ppb. The blowdown cation conductivity should be
below 2 u.S/cm and the sodium, chloride and sulphate concentrations in the blowdown
samples should remain below 100 ppb each, hi general, the heatup period should be used to
reduce impurity levels in the steam generator and prepare the secondary coolant system for
power operation. At least daily sampling of the feedwater and blowdown effluent are required
to maintain the above values (EPRI 1993c).
80
If copper alloy condenser material is used on the secondary side, the excess hydrazine
should be removed prior to heatup. Some of the hydrazine will thermally decompose to
ammonia during heatup, which at high concentrations will accelerate the copper alloy
corrosion and rapidly exhaust the condensate polisher resin. Also, the thermal decomposition
of the hydrazine will increase the pH and an upper limit of 9.2 prior to heatup is
recommended for systems with copper alloy material (EPRI 1993c).
Balance of Plant Corrosion. The principal method of controlling corrosion product

transport in the secondary systems of PWRs and the sludge buildup in their steam generators
is through pH control. EPRI recommends a room temperature feedwater pH between 8.8 and
9.2 for plants with copper alloys and above 9.3 for all ferrous plants (EPRI 1993c). hi
addition, the plant make-up and auxiliary feedwater should be deaerated and a chemistry
control programme should assure that the condensate/feedwater is clean, as discussed above.
Steam Generator Blowdown. The blowdown system should be able to accommodate a

continuous blowdown rate of 1% of the main steaming rate and a periodic, transient rate of
3-7% of the main steaming rate. The blowdown water should be processed through filters and
demineralizers and recycled to the condensate system to reduce the probability of oxygen
transport into the system.
Lancing. Lancing uses high-pressure jets to mechanically remove sludge from the
tubesheet face to alleviate IGSCC and IGA. By periodic sludge lancing, the depth of
accumulated sludge can be kept below the height necessary to cause dryout and concentration
of chemicals. Robotic inspection and lancing equipment, called CECIL (Consolidated Edison
Combined Inspection and Lancing System), has recently been developed. CECIL is equipped
with multidirectional pressurized water jets to remove sludge from around the tubes, grappling
tools to remove foreign objects, and a video camera to inspect the tube bundle as the work
progresses. This equipment allows close-up lancing of tenacious sludge deposits that cannot
be effectively removed by water jets from nozzles in the tube lane or tube bundle periphery.
Field tests at Indian Point Unit 2 in 1989 demonstrated that CECIL could remove about three
times the amount of sludge removed during conventional lancing, and that it removed a
significant amount of the hard sludge left after previous cleanings. CECIL has been used at
a number of plants in the USA, France, and Japan. Sludge lancing is conducted either every
or every other inspection period in France and Japan and every inspection period in Spain,
Switzerland, and Belgium.
Pressure pulse and water slap. The other mechanical cleaning processes, pressure pulse
and water slap, periodically release pressurized nitrogen at the bottom of the tube bundle. The
nitrogen produces upward movement of the water mass in the steam generator, thereby
dislodging deposits from the tube surfaces and from the tubesheet and tube-support plate
regions. The pressure pulse and water slap processes have been proven somewhat effective
in removing corrosion products. However, in contrast to chemical cleaning, the use of these
processes has only resulted in short-term improvements.
81
6. STEAM GENERATOR INSPECTION AND MONITORING REQUIREMENTS
AND TECHNOLOGIES
Steam generators are routinely inspected during plant outages, when their internal
structures become accessible to inspection equipment. This section identifies inspection and
monitoring requirements and techniques for steam generators, with emphasis on examining
the tubing, tubesheet, feedwater nozzle and shell. Tables VI, VII and Vffl list in-service
inspection methods used to detect damage from the various degradation mechanisms.
6.1. TUBING INSPECTION REQUIREMENTS
The probability and consequences of steam generator tube failures can be reduced
through appropriate and timely inspections. The steam generator tube inspection requirements
in the USA are discussed first because a number of countries with PWR and CANDU units
have used those requirements as a starting point for their own requirements. Tubing inspection
practices in Canada, the Czech Republic, France, Germany, Japan, Slovenia, Spain, Sweden,
and Switzerland are summarized in Table XXII and also discussed in this section. Tubing
inspection practices used in Russia and those recommended by EPRI are also discussed.
Tubing inspection requirements differ somewhat in these and other countries because:
• Different steam generator designs and materials and specific sites are susceptible to
different types of ageing degradation. Some types of degradation are easier to detect or
give rise to less severe safety consequences than other types of degradation.
• An appropriate level or steam generator and plant safety can only be maintained by a
suitable combination of inspection and acceptance (fitness-for-service) requirements.
Some countries have chosen to apply somewhat more conservative fitness-for-service
criteria and less inspection. Other countries have chosen less conservative fitness-for-
service criteria (thereby saving money on repairs) and more inspection.
• The frequency and scope of the inspections often increase as problems develop.
Complementary information concerning the fitness-for-service guidelines in various countries

and methods for assessing the residual life of the tubing is presented in Section 7.
6.1.1. Tiding Inspection Requirements in the USA
The requirements for the steam generator tubing inspections at US plants are included
in the plant Technical Specifications, which are prepared by the plant operator and approved
by the USNRC. Originally, those requirements generally followed the guidelines presented
in the USNRC's Regulatory Guide 1.83 (USNRC 1975). These guidelines are organized as
follows; access, equipment and procedures, baseline inspection, sample selection,
supplementary sampling, inspection intervals, acceptance limits, and corrective measures. In
summary, the steam generator should be designed with sufficient access to facilitate inspection
and plugging, eddy-current or equivalent equipment that is "sensitive enough to detect
imperfections 20% or more through the tube wall" should be used (unfortunately, reliable
82
TABLE XXn. SHAM GENERATOR TUBING INSPECTION GUIDELINES
Baseline Inspection Number of Tubes to be Inspected Inspection Intervals
•USA * All tubes pnor to • Fust inspection. 3% of the total steam • First inspection, 6-24 months
service and after any generator tubes at a unit • Subsequent inspections, 12-24 months
major change m • Subsequent inspections, see Table 23 • If less than 5% of inspected tubes with
secondary water indications and no defective lubes,
chemistry 40 months
• If more than 10% degraded and more than
1% defective, <20 months
Canada • 25% of the tubes • At least 10% of the tubes m one steam • Every S years
pnor to service generator per Ural
Czech Republic • All tubes pnor to service • At least 10% of the tubes m each steam • Every four years
generator must be inspected full length • Every four years
• Usually inspect all the tubes from the hot
collector and 50% of the tubes from the cold
collector
France • All tubes pnor to service • If susceptible tubing- all of the tubes are • Every outage for roll transition and
• All tubes every ten years inspected m die hot leg roll transition, tube small radius U-bend regions
(1st after 30 months) support plate and sludge pile regions, and the • Every other outage for TSP and
U-bend region of the first row in service sludge pile regions
with an appropriate probe
• If less susceptible tubmg Sample of tubes
inspected full length • Sample every two years
• All tubes in service with a previous defect
indication • Each outage
Germany • All tubes pnor to service • 10% of the tubes per steam generator per • Every four years all steam generators
inspection • Every two years, one half of the steam
generators
Japan * All tubes pnor to service • If no leakage and no defects 30% • If no leakage and no defects, every
• Insertion depth of anti- • If any leakage or defects 100% other year
vibration bars • If leakage or defects, every year
Slovenia • All tubes pnor to service • 100% using bobbin coil and all reported • Each refueling outage
indications, roll transitions and inner bends
with pancake owl
Spam • All tubes pnor to service • If susceptible tubing; 100% using bobbin • Each refueling outage
coil and all indications and roll transition
regions with rotating pancake coil
• If less susceptible tubing- 9 to 20%
Sueden • All tubes pnor to service • Random sample of 15-17% full length • Each year
• 100% hot leg tubesheet
• 20- 100% of other selected regions
Switzerland • All tubes after one year • If susceptible tubing:

of operation -inspect the hot leg side up through • Everyoutage
the U-bend region to the top tube
support plate on the cold side
-full inspection • Every three years
• If less susceptible tubing- random • Every three years
sample of 5.5% of all tubes
*If more than 10% of inspected tubes show indications, additional 3% m that steam generator and 3% in remaining steam generators
If more than 10% of second batch show indications, inspect additional 6% m area of indications.
83
detection of certain defect types at such a shallow depth is not within the state of the art), and
a baseline inspection of all tubes should be performed prior to service and after any major
secondary side water chemistry change.
Regulatory Guide 1.83 recommends that at least 3% of the tubes in each steam generator
be tested over their entire length during the first inspection, which should be performed after
six effective full power months but before 24 calendar months. Subsequent inspections should
not be less than 12 or more than 24 calendar months apart and may be limited to one steam
generator encompassing 3% of the total tubes at the plant. All non-plugged tubes with
previous indications (>20%) should be inspected. If any new indications are found (>20%)
or if previous indications exhibit growth (>10%) the remaining steam generators should be
inspected.
If more than 10% of the inspected tubes show indications (>20%) or one or more tubes
must be plugged (>40%), an additional 3% of the tubes must be inspected. If the additional
inspection indicates that more than 10% of the additionally inspected tubes have indications
or one or more of those tubes must be plugged, 6% more tubes should be inspected in each
steam generator. If two consecutive inspections result in less than 10% of the inspected tubes
with indications (>20%) and no further penetration of previous indications (<10%), the
inspection frequency should be extended to 40-month intervals. Unscheduled inspections
should be conducted in the event of primary-to-secondary coolant system leaks exceeding the
technical specifications or various design basis accidents (seismic, loss of coolant, main steam
or feedwater line breaks).
Regulatory Guide 1.83 was used as the basis for the steam generator inspection
requirements in the Technical Specifications for only a few years. By the early 1980s, the US
utilities were following the steam generator tube sample selection guidance in Table XXin
(USNRC 1981, Southern California Edison Co. 1982, Northern States Power Co. 1985,
Georgia Power Co. 1987, Commonwealth Edison Co. 1987). The tubes selected for each
inservice inspection include at least 3% of the total number of tubes in all the steam
generators at a unit and are selected randomly except:
(a) where experience in similar plants with similar water chemistry indicates critical areas
to be inspected, then at least 50% of the tubes inspected shall be from these critical
areas; and
(b) the first sample of tubes selected for each in-service inspection of each steam generator
generally includes all the tubes in service with previous indications greater than 20% of
the wall thickness; tubes in areas where experience has indicated potential problems; and
tubes adjacent to badly degraded tubes.
84
TABLE XXIII. STEAM GENERATOR TUBE INSPECTION REQUIREMENTS IN THE USA
1ST SAMPLE INSPECTION 2ND SAMPLE INSPECTION 3RD SAMPLE INSPECTION
Sample Size Result Action Required Result Action Required Result Action Required
A minimum of C-l None N.A. N.A. N.A. N.A.

S Tubes per
S.G. 02 Plug or sleeve defective tubes C-l None N.A. N.A.
and inspect additional 2S tubes
in this steam generator (S.G.) C-2 Plug or sleeve defective tubes C-l None
and inspect additional 4S tubes
in this S.G. C-2 Plug or sleeve defective tubes
C-3 Perform action for C-3 result of

first sample
C-3 Perform action for C-3 result N.A. N.A.
of first sample
C-3 Inspect all tubes in this S.G., All other None N.A. N.A.
plug or sleeve defective tubes S.G.s are
and inspect 2S tubes in each C-l
other S.G.
Some S.G.s Perform action for C-2 result N.A. N.A.
Notification to NRC pursuant to C-2 but no of second sample.
1|50.72(bX2) of 10 CFR Part 50. additional
S.G.s are
C-3
Additional Inspect all tubes in each S.G. N.A. N.A.

S.G. is C-3 and plug or sleeve defective
tubes. Notification to NRC
pursuant to 1]50.72(b)(2) of 10
CFR Part 50.
S = 3N/n% where N is the number of steam generators in the unit, and n is the number of steam generators inspected during an inspection.
oo
The results of each sample inspection are classified into one of the following three
categories:
Category Inspection Results

C-l Less than 5% of the total tubes inspected are degraded tubes and none of the
inspected tubes are defective.
C-2 One or more tubes, but not more than 1% of the total tubes inspected are
defective, or between 5% and 10% of the total tubes inspected are degraded
tubes.
C-3 More than 10% of the total tubes inspected are degraded tubes or more than
1% of the inspected tubes are defective.
Degraded tubes are tubes with indications greater than or equal to 20% of the nominal
wall thickness, but less than a defective tube, and which exhibit a defect with a greater than
10% additional wall thickness penetration since the last inspection. Defective tubes are tubes
with indications greater than or equal to the removal from service (plugging) or repair limit
which is often but not always 40% of the nominal wall thickness.
The first sample inspection defined in Table XXm requires a full end to end survey of
each of the tubes. The tubes selected as the second and third samples (if required) during each
in-service inspection may be subjected to a partial tube inspection provided:
(a) the tubes selected for these samples include the tubes from those areas of the tubesheet
array where the tubes with imperfections were previously found; and
(b) the inspections include those portions of the tubes where imperfections were previously
found
The in-service inspections shall be performed at intervals of not less than 12 nor more
than 24 calendar months after the previous inspection. If two consecutive inspections, not
including the pre-service inspection, result in all inspection results falling into the C-l
category or if two consecutive inspections demonstrate that previously observed degradation
has not continued and no additional degradation has occurred, the inspection interval may be
extended to a maximum of once per 40 months. If the results of the in-service inspection of
a steam generator conducted in accordance with Table XXIII at 40-month intervals fall in
Category C-3, the inspection frequency shall be increased to at least once per 20 months.
Additional, unscheduled in-service inspections shall be performed after the following
conditions: reactor-to-secondary tube leaks (not including leaks originating from tube-to-
tubesheet welds) in excess of the limits of the Technical Specifications; or a seismic
occurrence greater than the Operating Basis Earthquake; or a condition IV loss of coolant
accident requiring actuation of the engineered safety features; or a condition IV main steam
line or feedwater line break.
Two steam generators are inspected during the first outage at units with four steam
generators (4-loop Westinghouse-type plants) and then one steam generator is inspected
during the second and subsequent outages, unless additional inspections are required because
86
of extensive degradation as indicated in Table XXIII. Only one steam generator is inspected
during the first and subsequent outages at 2- and 3-loop Westinghouse-type plants and at
Combustion Engineering plants unless additional inspections are required because of extensive
degradation as indicated in Table XXIII.
If alternative fitness-for-service guidelines for ODSCC at tube support plates are used,
more comprehensive inspections must be required by the plant Technical Specifications
(USNRC 1995b). These include bobbin coil probe inspections of all the hot-leg tube support
plate intersections, all the cold-leg intersections down to the lowest cold-leg tube support plate
with known ODSCC, and 20% of the tubes full length. In addition, rotating pancake coil
inspections are required for all bobbin coil indications greater than 1.0 volt (19 mm diameter
tubes) or 2.0 volts (22 mm diameter tubes). Also, rotating pancake coil inspections are
required at all tube-to-tube support plate intersections with (a) interfering signals from copper
deposits, (b) dent signals greater than 5 volts, or (c) large mixed residuals.
6.1.2. Tubing Inspection Requirements in Canada
The most recent Canadian Standard, CAN/CSA N 285.4-94, requires an inspection of

25% of the tubes in each steam generator after the steam generator is first installed, but prior
to service. The initial in-service inspection sample and frequency is 10% of the tubes in one
steam generator of each unit every 5 years. The inspection sample expands when significant
degradation is detected as follows:
"additional inspection includes tubes surrounding those with indications, which do not
comply with the acceptance criteria, to bound the degradation. As a minimum the
additional inspection includes all tubes whose centre lines are located within a radius of
2.5 times the tube spacing from centre line of affected tube. Consideration shall be given
to additional inspection of similarly located tubes in other steam generators in the reactor
unit (and in other units for a multi-unit station)."
The Canadian Standard is somewhat vague about the numbers of similarly located tubes
in other steam generators that need to be inspected. Therefore, the Canadian regulatory
authorities expect the owner to develop specific proposals and submit them for unit restart.
The main reason for this is that the regulatory bodies consider the owner to bear the primary
responsibility for the safety and good performance of the reactors, and it is up to the owner
to define and defend a course of action for the inspections.
The Canadian Standard requires that alternative NDE techniques be used to detect defects
not readily detected by the standard bobbin coil eddy current inspection technique. Acceptable
alternative NDE techniques include the use of specialized eddy current probes such as the
transmit-receive Cecco probes and motorized rotating pancake coil probes. In addition, ultrasonic,
visual and profilometry inspection techniques should be used where appropriate. The choice of
an alternate inspection method depends on the type of degradation encountered and its location.
The Canadian Standard also requires that a section of one tube be removed from one
steam generator for metallurgical examination once every five years. This applies to the lead
unit in a multi-unit station. The results of these periodic destructive examinations are to be
used to calibrate the non-distinctive examination techniques.
87
The inspection discussed above describes the minimum requirements for Canadian
reactors. In practice, the plant operators exceed these requirements and prepare specific
programmes of inspection and assessment suited to the individual sites. Even a new station
such as Darlington exceeds these requirements in both extent and frequency. For example,
Darlington is a four unit station but it requires a tube to be removed from one steam generator
in each unit every two years. Also, the minimum inspection sample is 20% (instead of 10%)
of one steam generator. Also, ultrasonic and specialized eddy current probes are always used,
not just bobbin coil probes. This station has no active degradation mechanisms but is
following an aggressive inspection (and preventive) programme.
6.1.3. Tiding Inspection Requirements in the Czech Republic
The Czech regulatory agency requires a baseline inspection before operation and then
a minimum of 10% of the tubes in each steam generator inspected full length every four
years. (Each of the six steam generators at each WWER-440 unit are inspected every four
years.) However, recent practice has been to inspect all of the tubes from the hot collector
side and 50% of the tubes from the cold collector side.
6.1.4. Tubing Inspection Requirements in France
The French regulatory agency requires a baseline inspection before operation, periodic
inspections at least every two years, and complete inspections (presumably 100% of the tubes
full length) every ten years. The EDF guidelines for steam generators with susceptible tubing
(Alloy 600) require a 100% inspection of the hot leg roll transition region and the U-bends
of the first row in service every outage and 100% inspections of the hot leg tube support plate
and sludge pile regions every other outage, with follow-up inspections of indications during
the next outage. The roll transition and small radius U-bend inspections must be done with
rotating pancake coil eddy-current equipment. The tube support plate and sludge pile
inspections can be done with bobbin coil eddy-current equipment.
6.1.5. Tubing Inspection Requirements in Germany
The scope and frequency of tubing inspections in the Federal Republic of Germany are
specified in KTA 3201.4. Ten per cent of the tubes in each steam generator must be fully
inspected every four years and half the steam generators must be inspected every two years.
However, actual inspections have been more frequent and some Siemens/KWU steam
generators have been inspected every operating period over much of their life.
6.1.6. Tubing Inspection Requirements in Japan
The Japanese authority requires that 30% of tubes be inspected every other year when
a steam generator has had no leakage and no tube degradation. If any primary-to-secondary
coolant system leakage or any tube defects are detected, 100% of the tubes have to be
inspected each year over their full length. Before each inspection, the steam generator tubes
are subjected to a 13.8 MPa (2000 psi) differential pressure test to open tight cracks and make
them more detectable. Bobbin coil eddy current equipment is used above the tubesheet region.
Eight by one eddy current probes are used in the hot-leg tubesheet region in most steam
88
generators in order to detect circumferential degradation. Rotating pancake coil eddy current
equipment is used in the tubesheet region of one Japanese plant in order to detect pitting.
6.1.7. Tubing Inspection Requirements in Russia
Russian steam generator tube inspections are performed when leakage of the primary
coolant into the secondary coolant system is detected. All the tubes are inspected using
"visual and hydro-luminescent" methods. The secondary side is drained and pressurized with
gas and video cameras are placed inside the collectors to look for bubbles. Or a fluorescent
substance is added to the secondary water which is pressurized, the primary side is drained,
and the tube ends are inspected. Eddy current inspection is being introduced at some Russian
nuclear power plants. Primary-to-secondary leak rates are monitored using a 24Na device.
6.1.8. Tubing Inspection Requirements in Slovenia
Initially, the sampling procedure outlined in USNRC RG.1.83 was followed. However,
the condition of both steam generators triggered more extensive inspection. Current practice
is full length inspection of all tubes with bobbin coil probes. Additionally, all bends in Rows
1 and 2 and hot-leg transition zones are inspected with multifrequency rotating pancake coil
probes. Bobbin coil indications at the tube support plates are also re-inspected with
multifrequency rotating pancake coil probes for confirmation. The expanded tubes in the
preheater section (cold leg) are also inspected using rotating pancake coil probes. A complete
inspection is performed during each refueling outage.
All sleeves and the tube areas behind the sleeves are also inspected during each refueling
outage. Also, an ultrasonic baseline inspection was used to confirm the quality of the sleeve-
to-tube welds. I-coil and Plus-point eddy current probes have been employed for subsequent
examinations.
6.1.9. Tubing Inspection Requirements in Spain
All Spanish steam generators with susceptible material are inspected during each
refueling outage. All of the tubes are inspected over their full length using bobbin coil eddy
current equipment. All the hot-leg tubesheet areas and all the indications detected by the
bobbin coil are also inspected with rotating pancake coil eddy current equipment. Fewer tubes
are inspected in the Spanish steam generators with less susceptible material. For example,
only 20% of the thermally treated Alloy 600 tubes in the Westinghouse Model F steam
generators at one plant are inspected over their full length every outage with a bobbin coil,
plus a random sample are inspected with rotating pancake coil eddy current equipment (the
Model F has stainless steel quatrefoil support plates). In another plant with Alloy 800M
tubing, 9% of the tubes are inspected over their full length every outage.
6.1.10. Tubing Inspection Requirements in Sweden
In Sweden, a random sample of 15-17% of all tubes must be inspected full length using
bobbin coil eddy current equipment each year. In addition, an augmented inspection of
20-100% of all tubes at specific regions (roll transition, tube support plate, etc.) is performed.
89
The augmented inspections include 100% of the hot-leg tubesheet area. The Swedish
regulatory authority must witness the inspections.
6.1.11. Tubing Inspection Requirements in Switzeriand
The Swiss utility (NOK) practice is to carry out a 100% inspection of their newer steam
generators with Alloy 690TT tubing (which are not particularly susceptible to degradation)
after one year of operation. A random sample of 5.5% of all tubes must be inspected every
three years thereafter. In addition to a full inspection every three years, all the tubes in the
older steam generators with Alloy 600 tubing are inspected on the hot-leg side, and up
through the U-bend region to the sixth support plate on the cold leg side, every outage.
Multifrequency bobbin coil eddy current equipment is used for these inspections,
supplemented by rotating pancake coil inspections of the U-bends in Rows 1 and 2 as well
as indications within the tubesheet (including the roll transition region). The Swiss regulatory
authority must witness the inspections.
6.1.12. EPRI Tubing Inspection Recommendations
When the EPRI alternative fitaess-for-service guidelines for defects in the roll transition
region are used, EPRI recommends a rotating pancake coil eddy current inspection of all in-
service hot-leg tube expansion zones at each scheduled inspection outage. When the EPRI
alternative fitness-for-service guidelines for defects in the tube support plate regions are used,
EPRI recommends a bobbin coil eddy current inspection of all hot-leg tube support plate
intersections, and all cold leg tube support plate intersections down to the lowest tube support
plate with indications every outage. Supplemental rotating pancake coil inspections of a
sample of tubes with bobbin coil voltages less than the tube repair limit is also recommended
to characterize the defects.
It should be noted that the USNRC has not accepted the EPRI fitness-for-service
guidelines for defects in the roll transition region, and, therefore, the EPRI tubing inspection
recommendations for that defect type and location are currently not in the Technical
Specifications at the US plants. The USNRC has accepted certain alternative fitness-for-
service guidelines for ODSCC at tube support plates (see Section 7.2.2), but requires a
somewhat more extensive examination than recommended by EPRI, as discussed in Sections
6.1.1 and 7.2.2.
6.2. TUBING INSPECTION TECHNIQUES
6.2.1. Eddy Current Testing
Primary reliance is placed on eddy current testing because that technology works well
on thin-walled tubes of the sort used in PWR and CANDU steam generators and because of
the large number of tubes to be inspected. Two probes have been widely used: the standard
bobbin coil and the more sensitive multifrequency rotating pancake coil (MRPC). The great
length of tubing to be inspected, as much as 738 500 m per inspection, favours the speed of
the standard bobbin coil; more than 2770 m can be inspected per hour using a bobbin coil.
Supplemental examinations of areas of concern can be performed using the slower, but more
sensitive rotating pancake coil probes. However, the sizing capabilities for these inspection
90
methods are limited In some cases, the uncertainties in sizing of the defects are being
determined by comparing the inspection results obtained using eddy current, and sometimes
ultrasonic inspection, with metallographic examinations of pulled tubes. The advantages and
disadvantages of eddy current methods (and in some cases ultrasonics) in detecting different
degradation mechanisms in PWR steam generator tubing are briefly discussed here. The
CANDU experience is discussed separately at the end of this section.
Primary Water Stress Corrosion Cracking. Axial PWSCC in the roll transition region can
be detected with bobbin coil probes and circumferential PWSCC in the roll transition region
can be detected with multifrequency rotating pancake coils (3-coil probes). Other eddy current
inspection methods such as array probes, and Cecco-3 and Cecco-5 probes, and ultrasonic
inspection methods can also detect circumferential cracks. A rotating pancake coil can detect
a circumferential crack greater than 50% throughwall. However, any distortion in the
expansion transition region may mask the PWSCC cracks. Therefore, tubes with such
distortions should be examined with probes that can differentiate geometry variations from
the inside diameter cracks. Generally, eddy current inspection is performed while pulling the
probe through the tube. However, for the expansion transition region, it is recommended that
rotating pancake coil inspections be performed during insertion to eliminate the drop through
effect and improve the quality of the inspection results.
Cracks in the U-bend region are difficult to detect because of the eccentricity of the
probe while traveling in the bend. The bobbin coil probe appears to be able to detect axial
cracks in the U-bend regions only when the total number is beyond a certain threshold or the
cracks are long. Flexible bobbin coil probes seem to work best.
Axial cracks can be sized with rotating pancake coil probes. One estimate of the
accuracy of measuring the length of axial PWSCC using a rotating pancake coil probe is ±
1.5 mm. This estimate was obtained by comparing the MRPC length measurements with the
actual maximum length of about 60 axial cracks in the roll transition regions of six pulled
tubes from a Belgian PWR (Van Vyve and Hernalsteen 1991).
The eddy current (and ultrasonic inspection methods) capable of detecting circumferential
cracks cannot reliably size the length of these cracks. Currently, industry is working on
developing qualified sizing techniques. Eddy current inspection data using rotating pancake
coil probes and metallographic data for circumferential crack samples are being used to
develop these techniques. The samples were explosively expanded in a simulated tubesheet
and cracks were produced in an accelerated acid environment. This configuration simulates
a Combustion Engineering design steam generator. Worldwide pulled tube metallographic data
are also being used to develop the sizing techniques. The pulled tube data show that eddy
current inspections generally underestimate the actual arc length of a circumferential crack
as shown in Fig. 21. For example, an actual arc length of a circumferential crack was 360°
whereas the corresponding eddy current result was about 100 degrees. Enhanced analysis of
the rotating pancake coil inspection results can reduce this deficiency (Malinowski 1995).
Some industry efforts are also directed towards developing improved eddy current techniques
for sizing the maximum depths of circumferential cracks, which are needed for tube integrity
assessments (Dembek 1995).
91
360 -r
0)
w
u-
13
270 -•
o
TJ
LU
O) 180 -•
0)
Q
O)
c
90 -•
o
90 180 270 360

Arc Length, Deg., Metallurgical
FIG. 21. Comparison of actual an: length of circumferential cracks in pulled tubes with the
ones estimated using eddy current inspection. The data represents worldwide experience as
of 1992 (Malinowski 1995).
Eddy current techniques are not effective in detecting cracks with more complex
morphologies than a pure axial or circumferential orientation. In one Belgian plant, large axial
cracks in the roll transition region masked the response from a small circumferential crack
that was also present in the roll transition region; the rotating pancake coil probe did not
detect the circumferential crack.
In-service inspection results show that the crack propagation rate of axial cracks on the
inside surface in the roll-transition region can be estimated. However, crack propagation
kinetics for circumferential PWSCC are not yet well characterized.
Intergranular Attack. Intergranular attack is difficult to detect and characterize with eddy
current testing. This is supported by the eddy current inspection experience at the Trojan
plant, which indicated that the threshold at which intergranular attack can be detected reliably
with MRPC and bobbin coil probes is not well understood (USNRC 199la). Intergranular
attack results in a slow and progressive change in the electrical conductivity and magnetic
permeability of the material. Therefore, a bobbin coil probe in an absolute mode can provide
detection and some information about the extent of the attack, but a bobbin coil probe
configured in a differential mode is not sensitive to intergranular attack. Also the specialized
pancake coil probes, such as the rotating pancake coil probe (or ultrasonic probes) which are
sensitive to axial and circumferential cracks but insensitive to geometrical or magnetic
92
discontinuities, are not likely to detect intergranular attack (Roussel and Mgnot 1991). The
Trojan plant staff took a conservative position that MRPC indications at the tube support
plates may indicate intergranular attack greater than the 40 per cent plugging limit, even
where a bobbin coil probe did not detect damage.
Eddy current probes cannot accurately assess the propagation of intergranular attack
because these probes cannot characterize the damage.
Outside Diameter Stress Corrosion Cracking. The reliable detection and sizing of
ODSCC using eddy current testing is difficult because of the low signal-to-noise ratios
frequently exhibited by such cracks. ODSCC has been detected in the roll-transition and
explosively expanded zones, in tube-to-support plate crevices, especially in the steam
generators with drilled-hole support plates and in the freespan regions in steam generators
with heavy crud deposits. Intergranular attack, pitting, and denting are also sometimes
observed in these locations. Intergranular attack is associated with ODSCC and ODSCC may
also initiate in pits. In some plants, circumferential ODSCC appears to be associated with
significant denting at the support plates and was not detected by a rotating pancake coil probe.
Axial ODSCC cracks have been detected at the tube support plate intersections and freespan
locations, circumferential cracks have been detected at both the expansion transition region
and at the support plate intersections. Most of the ODSCC cracks are of short length (USNRC
1990).
An eddy current inspection with a bobbin coil probe may miss some axial ODSCC
cracks at tube support plate intersections. In one case, metallographic examination of a pulled
tube revealed axial cracks within two 30-degree-wide bands on opposite sides of the tube,
with the deepest one being 62 per cent throughwall. However, the previous field inspection
using a bobbin coil probe did not report these cracks using the plant voltage threshold criteria,
A rotating pancake coil probe, using a 2.92 mm (0.115-in.) diameter unshielded pancake coil,
may be used to detect axial ODSCC.
As discussed above, rotating pancake coil probes underestimate the length of

circumferential stress corrosion cracks, including ODSCC cracks. This fact is illustrated in
Fig. 22, which presents a comparison of eddy current measurements of the arc length of a
variety of circumferential cracks with the corresponding metallographic examination results
(Dembek 1995). For example, an actual arc length of a crack was about 270 degrees whereas
the corresponding eddy current measured arc length was 90 degrees. As mentioned above,
enhanced analysis of rotating pancake inspection results can provide better agreement between
the inspection and actual arc lengths (Malinowski 1995).
Use of multifrequency/multiparameter eddy current methods can help suppress the eddy
current response of unwanted parameters including the support plates, tube sheet, and tube
denting. As in the case of PWSCC, the eddy current techniques are not effective in detecting
ODSCC with more complex morphologies than a pure axial or circumferential orientation.
(Ultrasonic inspection methods are used at some plants to size the length and depth of the
ODSCC cracks.)
93
360 -r
0)
t: 270
3
o
>>
•a
•a
ill
o> 180
0)
Q
CO
c
0)
90 --
o
-H———————I———————I—
90 180 270 360
Arc Length, Deg., Metallurgical
FIG. 22. Comparison of actud arc length of circumferential ODSCC cracks in pulled tubes
with the ones estimated using eddy current inspection (Malinowski 1995).
Improved guidelines for detecting and sizing circumferential cracks using rotating
pancake coil probes are being developed. The detection limit for a circumferential ODSCC
crack in dents is about 50% of the wall thickness and a 50° arc length, or 100% of the wall
thickness and 23° arc length. Based on the metallographic data for pulled tubes, the Wextex
Owners Group has reported that the length of a circumferential crack in a Wextex expansion
region can be estimated with an accuracy of ± 39°. (Wextex expansion is an explosive
expansion of the tube over full depth of the tubesheet. This expansion method was used in
the Westinghouse-type steam generators during the early 1970s.) In January 1995, the EPRI
In-service Inspection Guidelines Committee co-ordinated a series of circumferential arc length
measurements using rotating pancake coil probes from several vendors plus Cecco-5 estimates
from Westinghouse. All these measurements estimated the arc lengths within ± 37° to 45°,
which is consistent with the Wextex Owners Group estimate (Malinowski 1995).
Pitting. The accuracy of an eddy current pit depth measurement is severely limited
because of the small size of the pits and because the pits are often filled with copper
containing corrosion products which have a high electric conductivity. Ontario Hydro
Technologies has developed a rotating ultrasonic inspection system for measurement of pit
depths in Monel 400 steam generator tubes. High ultrasonic frequencies, on the order of 25
MHZ, are used to obtain the required pit depth accuracy and water is used as a couplant. The
94
inspection system is capable of accurately measuring pit depths to ± 2% of the tube wall
thickness, which is equivalent to measuring a pit depth within m. The depth measurement
results on pulled tubes compared well with the metallographic results (Moles et al. 1994).
Denting. Bobbin coils are usually employed to detect and size most dents. (However,
rotating ultrasonic inspection probes can provide more accurate radial profiles of a dented
tube cross-section.) The denting growth can be considered as slow and its evolution as well
controlled. However, very small dents can initiate stress corrosion cracking but cannot be
detected with bobbin coil probes. French experts have calculated that dents as small as 20 to
30 microns are large enough to cause stresses capable of initiating stress corrosion cracking
in susceptible Alloy 600 tubing.
High-Cycle Fatigue. It is difficult to detect a high-cycle fatigue crack in a steam

generator tube because the initiation time for such a crack is quite long and the crack growth
is rapid Mitsubishi Heavy Industries (MHT) have evaluated use of pancake coils for detection
of circumferential fatigue cracks. Because the fatigue cracks are tight and rather straight, the
evaluation focused on how the detectability is affected by the width of the crack and the type
of the coil. The experimental results for the impedance of an artificial circumferential fatigue
crack, 2 * 10~3-mm wide and 50% throughwall, compared well with analytical results, and,
thus, validated the numerical analysis approach. The numerical solutions for different crack
widths showed that for very narrow circumferential flaws such as fatigue cracks, the
detectability of the crack is not affected by the crack width, and such cracks can be detected
by pancake coil type probes (Miyake et al. 1992).
Wear. Wear causes loss of material at the tube outside diameter. The shape of fretting
induced wear is determined by the contact area with the supporting structure and its length
is limited to the thickness of the supporting structure. Fretting induced wear is also limited
to some critical tubes, which makes inspection much easier. This damage is relatively easy
to detect and size with a bobbin coil probe, hi addition, field experience shows that the wear
rates tend to be low compared with the frequency of the in-service inspections.
Loose parts induced wear is generally limited to peripheral tubes and is also relatively
easy to detect when it is suspected However, the sizing of the affected area is less accurate
than that for the fretting induced wear because the shape of the wear is unpredictable. The
rate of the loose parts induced wear is also unpredictable and a long rapidly progressing wear
scar may lead to tube rupture in less than a fuel cycle.
Erosion-corrosion. Erosion-corrosion is detected by conventional bobbin coil probes

during normal in-service inspections of steam generators, hi the majority of cases, indications
are detected at least one or two inspection intervals prior to the wall reductions reaching 40%
of the initial wall thickness. Hence, it is concluded that there is some forewarning of damage
by this degradation mechanism before it reaches a state where reliability is affected.
Corrosion-fatigue. The metal loss associated with high cycle corrosion assisted fatigue
can be detected by eddy current methods during the normal in-service inspection of a steam
generator. However, some cracks are difficult to pick up due to their close proximity to the
tubesheet and tube support plates. It is not clear from the limited data whether a gradual
95
degradation can be detected before sufficient degradation has occurred to require removal of
the tube from service.
Wastage. There is a general consensus that wastage can be accurately detected and sized
using a bobbin coil probe when the wall loss is larger than 10% to 20%. Operating experience
shows that the propagation rate of wastage is compatible with the frequency of in-service
inspection.
Summary. Inspection of the steam generator tubes is critical to the safe and economical
operation of nuclear power plants. Eddy current inspections using bobbin coil probes were fast
and effective in detecting and sizing the degradation that took place in early steam generators.
However, newer forms of degradation have appeared in recent years that require development
of more sophisticated inspection tools. Often, different characteristics of the damage require
different types of inspection tools.
The safety significance of the uncertainties in the eddy current technologies varies. For
example, most PWSCC can be detected and then characterized well enough to make a repair
decision with the available probes. Detection of IGA patches with standard multifrequency
bobbin coil probes is poor. However, this degradation is generally not deep enough to have
any safety significance. Occasionally, there have been problems detecting ODSCC and three
tube ruptures have occurred to date because of undetected ODSCC. Pitting is also very
difficult to detect but is not expected to cause much leakage or contribute to a tube rupture.
Most dents are detected relatively easily with conventional bobbin coil probes. However, very
small dents can initiate cracking but cannot be detected with bobbin coil probes. In-service
inspection is not an effective approach for preventing high-cycle fatigue ruptures because of
the rapid crack growth rates; two such ruptures have occurred to date. Wear and fretting
damage is characterized by significant quantities of metal loss over an extensive area and as
such is easily recognized by conventional eddy current methods. However, loose parts wear
can sometimes occur very rapidly. Wastage and erosion-corrosion (and probably high-cycle
corrosion assisted fatigue) can be detected with eddy current technologies.
The primary method of tube inspection used in the CANDU units is also the standard
multifrequency bobbin coil eddy current method. Inspection probes and equipment have had
to be adapted to the following CANDU steam generator characteristics: smaller tube
diameters, reduced size of the primary head and the presence of a heavy magnetite layer on
the inside surface of the tubes which abrades the probes. The circumferential stress corrosion
cracking experienced at some units required the development of more sensitive and reliable
eddy current probes. The special multichannel transmit-receive probes, developed by the
Chalk River National Laboratory, are capable of detecting circumferential cracks >60%
throughwall.
Inspection of Monel tubes with eddy current probes posed some difficulties due to the
slightly ferromagnetic properties of the material. This makes it necessary to magnetically
saturate the tube material in order to detect flaws. This is done by using an eddy current
probe with a very strong permanent magnet built into the probe.
96
6.2.2. Ultrasonic and Other Inspection Methods
Ultrasonic testing is a volumetric nondestructive method for in-service inspection of

components. Ultrasonic methods which can detect pits and circumferential cracks in the
presence of axial cracks are being developed. This section discusses several ultrasonic
techniques used in tube inspection.
Electromagnetic Acoustic Transducer (EMAT). This method addresses ultrasonic

inspection of steam generator tubing using an electromagnetic acoustic transducer [Thompson
and Elsley 1983]. It was developed to detect flaws in certain areas of tubes where the
conventional single frequency, differential coil eddy current probe has not been entirely
satisfactory, including in circumferential cracks, defects at dents or support plates, and defects
in U-bends.
The EMAT system has a good detection capability for circumferential cracks and other
defects that provide a fairly wide and sharp circumferential oriented cross section (e.g., dented
areas); however, the system has difficulty detecting flaws with small cross sections, i.e.,
cracks that are tight. For example, the system can inspect U-bends but does not always detect
the axial cracks found in U-bends. hi addition, the system has limited defect depth-sizing
capability.
Pulse-Echo UltrasoundforTube-to-Support Plate Gap Measurement. This equipment uses

an ultrasonic technique for determining the condition of the gap between steam generator
tubes and support plates, which allows monitoring of corrosion product buildup that might
lead to denting. It also can determine the efficiency of chemical cleaning in removing this
buildup.
Optical Profilometry. The optical profilometer has been tested successfully under
laboratory conditions for tube dent measurement [Oberg 1983]. It can measure inside tubing
profiles in the range of radii from 8.13 mm to 10.16 mm, with an average calibration error
of 0.13-0.20 mm on the nominal inner diameter and i<3.15 mm on dents. Preliminary
investigations suggest that the calibration error is most probably caused by variations in the
interior surface finish.
Pitting. An automated ultrasonic system has been used at some CANDU units to detect
shallow tube pits. This system uses high frequency ultrasonics (50-100 MHZ) and is capable
of detecting pit sizes »5% throughwall.
6.2.3. Destructive Testing
Eddy current inspection (and ultrasonic examination) techniques and procedures can be
qualified by removing previously inspected tubes from an operating steam generator and
examining the defect indications in a laboratory. Appropriate destructive examinations of
so-called "pulled tubes" will not only quantify the defect indications but also provide
considerable information about the degradation mechanisms. Specifically, pulled tubes can be
used to determine if secondary-side corrosion defects are acid or caustic induced and identify
species associated with the chemical attack. Pulled tubes can also be used to determine leak
97
rates (with mixed success to date) and burst pressures, information which is useful for
assessing tube integrity. Furthermore, inspection of pulled tubes provides an opportunity to
look for any incipient problems. The tubing selected for destructive examination should
obviously have some defects, such as a roll transition indication, and must be accessible.
6.3. MONITORING LEAKAGE FROM TUBING
On-line monitoring of nitrogen-16 in the steam lines can point to rapidly increasing
primary-to-secondary leaks associated with, for example, high-cycle fatigue cracking in the
U-bend region. Nitrogen-13 is produced in the primary water as it passes through the reactor
core. It will be present as nitrogen-16 in the secondary system only if there is a primary-to-
secondary leak. Nitrogen-16 does not accumulate in the secondary system because its half-life
is only 7.35 seconds. Therefore, its presence provides a good measure of the current primary-
to-secondary leak rate with a very rapid response time. (The threshold value is less than 1
litre per hour if the plant is at a nominal load.)
Helium is used to search for air leaks after the unit reaches partial load. A load of at
least 20% is necessary for efficient leak testing.
6.4. FEEDWATER NOZZLE INSPECTION
The ASME Code, Section XI, provides the in-service inspection requirements for the
steam generator shell, feedwater nozzle, and the adjacent feedwater piping. There are no
inspection requirements for the feedring, J-tubes, or thermal sleeves, although those
components have experienced erosion-corrosion damage (wall thinning) in the field. In
accordance with the Code, the ASME requirements include surface and volumetric
examination which focus mainly on the feedwater piping welds and base metal immediately
adjacent to the welds, feedwater nozzle blend radius, and steam generator shell girth welds.
Radiographic and ultrasonic inspections have been used for this purpose. However, thermal
fatigue cracks, particularly in the base metal away from the weld, are not always detected
with an ASME examination. This section focuses on the inspection of the nozzle-to-pipe
welds and adjacent piping welds because of some recent cracking incidents at these sites.
ASME Section XIIn-service Inspection Requirements. The current pre-service and in-
service inspection requirements include a volumetric examination of the inner 1/3 volume of
the piping welds and adjacent base metal for a distance of 6 mm from the edge of the weld
crown and a surface examination of the outside diameter surface of the weld and 13 mm of
the adjacent base metal. The Code contains similar inspection requirements for the feedwater
nozzle-to-vessel welds. The inspection requirements for the shell welds such as the girth weld
include an examination of the entire volume. In addition to the welds, the Code also requires
a volumetric examination of nozzle inside blend radii. For multiple vessels of similar design,
such as the steam generators, the examinations may be limited to the nozzles of one vessel
or the equivalent of one vessel distributed among the vessels.
Although pre-service and in-service inspections are currently required across the nuclear
industry, they were not part of ASME Section XI Code prior to the Winter 1972 Addenda of
the 1971 edition. Thus, when cracking was discovered in the D. C. Cook nozzles in 1979,
98
many plants had not performed pre-service or in-service examinations of the feedwater
system. This lack of examination results was significant in that many plants did not have base
line ultrasonic examination data for the examiners to compare in discriminating geometric
reflectors, such as the counterbore and weld root, from service-induced defects. This is
especially important in the case of feedwater piping cracking, which has generally initiated
at geometric discontinuities such as the counterbore comer. For plants that are performing
examinations of welds in accordance with latter editions of ASME Section XI, the
examination volume may not extend far enough to include the discontinuity at the counterbore
corner. Some utilities are now including the examination of the counterbore corners in the in-
service inspection of the feedwater system.
Some recent events have illustrated the potential weaknesses in the ASME in-service
inspection requirements. In March 1992, through-wall cracking was discovered in a PWR
feedwater nozzle-to-transition piece weld at one US plant. Subsequent radiographic
examinations revealed that several nozzle-to-transition welds contained significant cracking.
All of these welds had been previously examined ultrasonically. Further investigation revealed
that the ultrasonic examinations were conducted using the minimal Code requirements and
that the indications had been incorrectly identified as root geometry. No supplemental or
enhanced techniques were used to verify that the indications were not cracks. As a result of
these failures, the plant operator upgraded their ultrasonic procedures and expanded the
examination volume for welds in piping and nozzle locations subjected to thermal
stratification. The expanded volume includes the weld plus the adjacent base metal for a
distance of two wall thicknesses. The plant operator also incorporated into their procedures
a number of enhanced ultrasonic techniques to aid in the evaluation of detected indications.
In addition, the in-service inspection personnel were provided enhanced training using the
removed pieces of the damaged feedwater piping and nozzle.
Another incident involving misinterpretation of crack indications occurred at another US

plant. These cracks were oversized by ultrasonic examination, which resulted in removal of
the affected weld. These same cracks were not detectable with radiographic examination.
Subsequent metallurgical evaluation results indicated that the ultrasonic examination had
oversized a shallow crack by a factor of ten owing to inclusions in the weld area. The plant
operator concluded that the ASME Code examinations were not adequate for small thermal
fatigue cracks and that enhanced ultrasonic techniques, such as tip-diffraction and creeping
wave techniques and automated scanning, were necessary to improve reliability, accuracy, and
repeatability.
Improved In-service Inspection Methods for Thermal Fatigue Cracks. Radiographic or

ultrasonic testing may be used for detection of cracking at the inside surface of the piping.
Each method has its own advantages and disadvantages, but is capable of providing
complementary information on the condition of a weld and the adjacent base metal.
The inherent advantage of radiography is that testing can be performed through the
insulation and a permanent record is obtained, which can be compared with the results of
future examinations. The resulting image can also be used to characterize the weld geometry.
Many of the disadvantages of radiography stem from convenience factors. These include
radiological controls that may interfere with critical path activities, interference from
99
contaminated and irradiated components, and access to the inside surface (which is available
only at those plants with gamma plugs installed in the pipe wall adjacent to the feedwater
nozzle). Otherwise, double wall techniques must be used, which have reduced sensitivity. The
final consideration is that radiographic examination is arguably a less sensitive method for
crack detection compared with ultrasonic examination. Although radiographic examination is
sensitive to defects that are volumetric in nature (e.g., wall thinning, slag inclusions, etc.), the
density difference caused by a crack may be insufficient for detection if the orientation of the
crack is not parallel to the gamma or X ray.
Although ultrasonic examination can be sensitive and capable of detecting many types,
sizes, and orientations of cracking, manual ultrasonic examinations have had two inherent
disadvantages: reliance on the inspectors' ability and judgement, and the lack of a permanent
record. For the feedwater piping, these problems have resulted in inconsistent results, miscalls
of both cracks and weld geometry (owing to the lack of baseline data), and a general lack of
confidence in ultrasonic examination. These inconsistencies and miscalls have led to the
development of enhanced inspection techniques.
Tip-diffraction techniques are widely used for crack depth sizing. The time-of-flight-
dififraction technique is one of the tip-diffraction techniques that has been developed in recent
years (Pers-Anderson 1993). The time-of-flight diffraction signals associated with different
crack configurations are illustrated in Fig. 23. As shown in Fig. 23(a), two signals are present
in the absence of a crack: a direct lateral wave signal and a signal reflected from the
backwall. Diffraction occurs when the incoming sound beam impinges upon a finite planar
reflector such as a crack. The diffracted sound energy from the crack tip acts as a point
source and radiates a sound wave to the receiving transducer. The time of arrival of this
signal can then be used to pinpoint the tip of the crack and determine the crack depth. Figure
23(b) illustrates such a diffracted signal produced by the tip of a surface crack; note the
presence of a backwall reflection signal and the absence of a lateral wave signal. The
presence of a surface crack at the inside diameter will cause the loss of the backwall
reflection signal, but a lateral wave and a diffracted signal from the crack tip are present, as
shown in Fig. 23(c). All signals will be present for an embedded crack, as shown in Fig.
23(d). This approach provides a means of sizing, locating, and verifying the extent of the
crack, but could be hindered by weld geometry on the outside and inside surfaces, which
could cause a loss of surface contact and/or loss of backwall and lateral wave signals for
reasons other than the presence of a crack.
The improvement in sizing technology using the time-of-flight techniques is largely the
result of the data published from the Programme for Inspection of Steel Components (PISQ
Round Robin Tests (the PISC n Project), which showed that amplitude based methods were
unreliable for deteimining the through-wall extent of a crack (Cowfer 1989). Since that time,
numerous transducer configurations (e.g., tandem, dual) have been developed to optimize the
response from the crack tip and other portions of the crack to aid in crack sizing.
An approach developed for the detection of intergranular stress corrosion cracking in

BWR recirculation piping that can be effective for the detection of thermal fatigue is the use
of an inside diameter creeping wave and the related mode conversion techniques. This family
of techniques has gained wide acceptance in the nuclear industry because of its high
sensitivity to small connected surface flaws at the inside surface (Brook 1986).
100
;-<-PCS
(a)
-v-
PCS a probe center separation Signal 1 Lateral wave
Signal 4 Backwall
3
(b) -V-
Signal 3 Lower crack tip

Signal 4 Backwall
1 2
(c)
Signal 7 Lateral wave

Signal 2 Upper crack tip
1 2 3
(d) -V^r-Y-4r
Signal 1 Lateral wave

Signal 2 Upper crack tip
Signal 3 Lower crack tip
Signal 4 BackwaJI
FIG. 23. Examples of time-of-flight diffraction (TOFD) signals (Pers-Anderson 1993).

Copyright TRC; used with permission.
Another improvement is the use of automated inspection equipment to collect and store
ultrasonic data. Automated scanning of feedwater nozzles was performed at San Onofre Unit
3 using the Introspect/98 volumetric inspection system (Mostafa and Ramsey 1994). Using
computer processing, 3-dimensional imaging of the data facilitated flaw characterization and
discrimination of geometrical reflectors on the inside surface. Such enhanced evaluations are
becoming routine and are possible with many modem scanning systems.
101
In the past, the primary constraint for automated ultrasonic inspection was collecting and
manipulating large quantities of data. However, the continued evolution of computers has
allowed for storage of gigabytes of data, and the speed to collect and manipulate the data.
Modern computers have also provided a more efficient means of processing data and
integrating processing with special search units. A good example of this is the TestPro/FATS
system (Bisbee 1994). FATS (focused array transducer system) is an extension of the phased
array techniques, which allow the beam to be focused electronically to the area of interest.
This method reduces beam spread and allows the beam to focus on the crack opening to
enhance detection, or focus on the crack tip to improve sizing accuracy.
One of the problems with performing ultrasonic examinations on piping welds is

discriminating weld geometry from cracks. Therefore, an accurate representation of the weld
geometry is important. Owing to the variation in weld geometry from weld to weld and to the
inaccuracies of as-built drawings, weld profiles for each weld examined are a necessity. One
method of obtaining this is by plotting thicknesses obtained by ultrasonic thickness
measurements. An alternate method that has been used in Japan to visualize weld geometry
is the use of computed tomography imaging of radiographic data (Maeda and Yagawa 1991).
The advantage of this approach is that an accurate cross section of any section of the weld,
including the reentrant corner of the counterbore, can be obtained. The disadvantage is, of
course, the expense of performing such an examination.
6.5. MONITORING FATIGUE DAMAGE TO FEEDWATER NOZZLES
This section briefly describes fatigue monitoring programmes offered by vendors in

Europe, Japan and the USA. It explains how some of the fatigue monitoring approaches have
been implemented, with emphasis on feedwater nozzles.
US Approach. The nuclear steam supply vendors and EPRI have developed several
systems for fatigue monitoring at critical sites in nuclear power plants. One such system is
described below. In 1984, EPRI began developing this system through a contract with
Structural Integrity Associates. The system includes a PC based on-line monitoring system
called FatiguePro that collects existing plant instrumentation data and then uses a Green's
temperature-stress function and an ordered overall range counting method to process the data
[Ware 1993]. The plant instrumentation data such as pressure, temperature, and flow rate are
used with the Green's influence function to determine stress versus time at the critical
locations. The ordered overall range cycle-counting method is used to develop a
stress/frequency spectrum from a measured stress history. The stress/frequency spectrum is
combined with the ASME Code S-N curve to determine the fatigue usage. Improvements to
the software, based on actual plant experience and fine tuning, include a correlation of
thermal stratification interface level as a function of feedwater flow. FatiguePro has been
installed in at least two Westinghouse four-loop plants.
Siemens-KWU Strategy. A Fatigue Monitoring System (FAMOS) has been developed

by Siemens-KWU to quantify fatigue usage. The system uses both a global and local
monitoring approach. Global plant process parameters have proven sufficient for most
components such as the steam generator shell. However, KWU determined that global
monitoring was not sufficient for the feedwater nozzle, so additional thermocouples were
102
mounted around the outside circumference of the feedwater nozzle. Data are obtained from
these thermocouples and the corresponding heat conduction problem is solved to determine
the stress and temperature fields. Once the stresses are known, a standard scheme for cycle
counting (rainflow cycle counting) and fatigue usage calculations is carried out. KWU
reported that thermocouple readings from the outside surface provided accurate temperature
estimates for the inside surface for wall thicknesses up to about 50 mm. Most German nuclear
power plants use some or all aspects of this fatigue monitoring strategy [Ware 1993].
French Strategy. Electricite de France was perhaps the first to implement fatigue
monitoring in nuclear plants and has one of the most mature programmes in the world. The
first EOF programme, the Transient Monitoring and Logging Procedure, was initiated with
the commissioning of its earliest plants (in 1977) and consisted of a bookkeeping method. The
actual transients were logged and checked, case by case, to ensure they were not more severe
than the design basis.
Even though the monitoring and logging procedure worked for many of the plant
components, EDF realized that advanced fatigue monitoring systems were required for more
critical locations. Consequently, EDF developed a system called the Fatiguemeter, which uses
plant instrumentation for measuring temperature, pressure, flow rate, and valve position (open
or closed). An influence function approach for estimating thermal stresses, rainflow counting,
and Miner's linear damage rule were then used to calculate fatigue at various locations.
Although the manual bookkeeping used as a part of the Transient Monitoring and Logging
Procedure reduced the cumulative usage factor substantially, the use of the Fatiguemeter
system resulted in an even further reduction in the usage factor (Bimont and Cordier 1989).
One of the applications of the Fatiguemeter was monitoring of the steam generator feedwater
nozzle at a 900-MW PWR plant.
After more than ten years of experience with the Transient Monitoring and Logging
Procedure and six years with the Fatiguemeter, EDF, in collaboration with Framatome, has
developed an integrated system called SYSFAC (SYsteme de Surveillance en FAtigue de la
Chaudiere). This system incorporates the lessons learned from the previous two approaches,
the logging system to review overall plant fatigue usage and a more detailed approach using
the Fatiguemeter at key locations (Bimont and Cordier 1989). The EDF standardized plant
design is helpful in developing this system because there are few plant-to-plant variations that
have to be considered, and the lessons learned from one plant can be directly applied to
others. SYSFAC is undergoing the final phase of testing and will apparently be installed at
a few French plants (Kergoat et al. 1994).
Japanese Strategy. Mitsubishi Heavy Industries Ltd has developed an on-line fatigue
monitoring system for nuclear power plant components (Masamori et al. 1988). The project
included development of surface temperature measurement instrumentation, of simplified
stress analyses and a prototype system Several locations, including the PWR feedwater
nozzle, are being considered for local monitoring of transients causing fatigue damage.
Recently, Sakai et al. (1995) developed an advanced transient and fatigue usage
monitoring system the fatigue monitoring system of the Japanese PWR Group (FAMS). The
system is based on the aforementioned Green's function approach but has an enhanced feature
103
for monitoring fatigue usage caused by thermal stratification. The heat transfer coefficient at
the piping inside surface is a function of temperature and flow velocity. To account for the
changes in the heat transfer coefficient during thermal stratification, the system software has
a Green's function database using the estimated heat transfer coefficient on the inside surface.
The software has been verified using measured data obtained from a PWR plant and three-
dimensional finite element analyses. Sakai et al. report that the first application of the FAMS
system to a Japanese plant will be in the near future.
104
HINESSPOR
SERVICE GUIDELINES
7.1. TUBING REPAIR CRITERIA
Repair or removal from service (plugging) of excessively damaged steam generator

tubing is necessary to prevent:
• single or multiple tube ruptures

• excessive primary-to-secondary leakage.
However, a continuing issue has been exactly what constitutes excessive damage and which
degraded tubes are or are not still fit for service. Some of the earliest guidance on this subject
was published in the U.S. Code of Federal Regulations and in the American Society of
Mechanical Engineers (ASME) Pressure Vessel and Boiler Code and is discussed in Section
7.2.1 below. The ASME code states that for U-tube steam generators, the allowable outside
diameter flaw shall be less than 40% of the tube wall. This criterion was initially implemented
in most countries with PWR or CANDU plants. However, alternative criteria are allowed by
the ASME code if accepted by the regulatory authority and USNRC Regulatory Guide 1.121
provides guidance on how to develop alternative criteria. Essentially, four items must be
addressed:
• the maximum (critical) size of a defect which ensures stability of the damaged tube
(analytical and experimental verification);
• the propagation rate of the defect until the next inspection;
• the ability of the inspection methods to detect defects of a critical size;
• the accuracy of the inspection methods to size defects of a critical size.
hi recent years, a number of countries have found the original ASME criterion overly
conservative and inflexible and have developed revised or new fitness-for-service criteria,
often in conjunction with revised inspection requirements. Although the new fitness-for-
service criteria used in most countries follow the general guidance contained in Regulatory
Guide 1.121, there are substantial differences in implementation. However, the currently
implemented repair criteria can be grouped into two families: generic and defect type and
location specific criteria. Both these types of fitness-for-service criteria are briefly introduced
below and then discussed in more detail in Section 7.2.
7.1.1. Generic Fitness-For-Seivice Criteria
No flaws. The simplest, most straightforward, and most conservative generic approach
is to define a minimum detection threshold, inspect all the tubes on a regular basis, and
remove from service or repair any tubes with indications above the noise level. This implies,
of course, that there will be no leakage. (Should any leakage start, the plant will immediately
be shut down and the tubes inspected.) However, this approach provides little or no incentive
to improve the inspection and leak detection methods.
Wall thickness. The most widely implemented fitness-for-service criterion is a minimum

wall thickness criterion (either the value specified in the ASME code or some other value).
105
The minimum wall thickness value is determined by assuming uniform wall thinning around
the circumference of the tube and calculating a wall thickness which will sustain all
postulated loads with appropriate margin. Generally, a plastic load limit analysis is performed
with margins against tube burst of 3 and 1.43 for normal and accident conditions,
respectively. Leak rate calculations are not required since throughwall defects are not
expected. A minimum wall thickness criterion works well for degradation mechanisms that
remove considerable material such as loose parts wear, wastage, etc. However, a minimum
wall thickness criterion can be overly conservative and costly for small defects such as pitting,
axial ODSCC within the tube support plates, etc.
7.1.2. Defect Type and Location Specific Repair Criteria
The occurrence in recent years of new types of tube degradation such as PWSCC within
the tubesheet or axial ODSCC within the support plates initiated the development in some
countries of defect type and location specific repair criteria. These criteria were developed to
reduce the extent of the steam generator repair or plugging work without sacrificing plant
safety by reducing the unnecessary conservatisms of the generic criteria. This was done by
taking into account specific defect and location characteristics which may reduce the chances
of tube rupture or leakage. To date, four broad groups of defect specific repair criteria are in
use.
P and F* Criteria Tubes with flaws in the region where the tube has been expanded
against the tubesheet will not burst and probably will not leak. Therefore, criteria were
developed specifically for full tubesheet depth expanded tubes, which allow tubes with flaws
in the tubesheet region to remain in service without repair, regardless of defect size. However,
the flaws must be some distance below the top of the tubesheet or bottom of the roll
transition, whichever is lower, so as to prevent pull out of the damaged tube should it separate
at the flaw. The F* distance for full depth rolled steam generators is typically 38 to 51 mm.
(The exact F* distance is established by considering the length of roll expansion needed to
resist the tube pull out forces.) The P* distance is typically about 32 to 38 mm. It is
established by considering the ability of other tubes to prevent tube pull out (Gorman et al.
1994). The tube sheet thickness is usually between 525 and 610 mm.
Crack length criteria for axial PWSCC in the residual stress dominated expansion
transition zones. This type of repair limit was originally developed and implemented in some
European countries (France, Belgium, Spain, Sweden, Slovenia). Axial cracks located close
to the top of the tubesheet and shorter than about 10 mm (3/4" tubes) or 13 mm (7/8" tubes)
may remain in service even if they are throughwall. Implementation requires special
inspection techniques which are able to detect and size the length of the axial cracks and,
depending upon the degree of the degradation, up to 100% yearly inspections.
The main underlying assumption is that Alloy 600 is very ductile. Therefore, reasonably
short throughwall axial cracks exhibit slow propagation (typically about a mm/year) and do
not tend to result in catastrophic tube failure. Rather simple analytical models (e.g., Erdogan
1976, for application see Flesch and Cochet, 1990) complement the experimental results well
and enable reliable predictions of critical crack lengths. Crack growth predictions are
estimated on the basis of statistical analyses of consecutive inspection results. The accuracy
106
of the inspection methods is determined using the results of metallographic examinations of
pulled-out tubes.
Leakage from tubes with various size cracks has been measured in the laboratory (Flesch
and Cochet, 1990). It was shown that the leak rate through a single throughwall crack of
about critical length is less than 70 L/h (0.3 gpm). Therefore, reduced operational leak rate
limits (below 70 L/h per steam generator) and on-line leak rate monitoring (such as nitrogen-
16) were carried out as an additional safety precaution.
The Swedish application of this criterion has an additional very interesting feature. The
final value of the repair limit is chosen on the basis of probabilistic safety assessment
analysis. The acceptable conditional tube rupture probability, given a steam line break, was
set to 1%, which implies an acceptably low core melt frequency (German, 1994, and
references therein).
Leak before risk of break criteria for axial PWSCC. This approach is of French origin
and is very similar to the crack length criteria. In the early implementation stage, leak
detection was considered to be as reliable as tube inspection. Only samples of tubes were
therefore inspected while the non-inspected, and possibly nearly critical, defects were
expected to be reliably detected by nitrogen-16 on-line leak monitoring. However, some of
the long throughwall cracks are rather leaktight, which can cause unreliable predictions of the
leak rates. The current tendency is therefore to put increasing weight on the use of inspections
and use leak detection as an additional safety feature.
Voltage criteria for ODSCC at the tube support plates. The very complex morphology
of ODSCC forced the industry to a completely statistical approach. The signal amplitude of
the bobbin coil eddy current testing inspection method was taken as the measure of defect
severity. Based on degraded pulled-out rubes and specimens prepared in the laboratory, two
correlations were developed: (1) bobbin coil signal amplitude versus tube burst pressure and
(2) bobbin coil signal versus leak rate (individual defect in a tube). The burst pressure
correlation, together with allowances for defect progression and inspection uncertainties, is
used to define the structural repair limit in the first step. The leak rate correlation, together
with the recent population of defects in the steam generator under consideration and
allowances for defect progression and inspection uncertainties, then gives an estimate of total
leak rate during postulated accident conditions (e.g., steam line break). Should the total leak
rate exceed the predefined acceptable value, the plant operator has the following options:
- repair or remove from service (plug) additional tubes,

- lower the reactor coolant system activity limits, or
- reduce the time between inspections.
Thus, the repair limit may depend on the condition of the steam generator, the growth rate
of the defects, the coolant activity levels, and other factors, and may be updated at each
inspection and repair campaign.
Note that the voltage criterion is not based on a mechanistic description or modelling
of the defect in contrast to other criteria. Rather, a simple correlation between a selected
107
o
oo
TABLE XXIV. GENERAL INFORMATION ON REPAIR CRITERIA CURRENTLY IMPLEMENTED
Generic criteria Defect specific criteria
Paianietcr No leak, no flaw Wall thickness P, F Crack Length Leak before risk of break Voltage criteria
(full de|>th rolled tubes) (axial PWSCQ (axial PV\SCQ (ODSCC at TSP)
Repair Unit
Measure of defect seventy N/A Remaining wall thickness Distance from the top of Crack length Crack length and leak rate Signal anphtude [V]
[%] TS
Typical allowable value Detection threshold 40-50% >38-76mm <IOmm(3/4") <l 3 mm (7/8") 1-15 V
<I3 mm (7/8")
Inspection
Method Adequate to detect Bobbin Bobbin MRPC, UT MRPC Bobbin, confirmed by

degradation MRPC
Extent (mm) 100% 3% 100% (PS) 100% (top of TS) 12% (top of TS) 100% bobbin2
Frequency (max ) 1 year 3 years 1 year 1-2 years 1 ycai 1 year
Leak through defects
Required monitoring Sampling, on-line Sampling Sampling On-line On-line On-line
Impact on repair limit None None None None None Yes, increases number of
plugged tubes
Generic criteria Defect specific criteria
Rwtmeter No leak, no flaw Wall thickness P, F Crack Length Leak before risk of break Voltage criteria
(full depth rolled tubes) (axial PWSOQ (axial PV\5OQ (ODSCC at TSP)
Typical operational limit per 0 38 l/mm(l gpm) 3 8 I/mm (1 gpm) <1 9 I/mm (05 <0 4 I/mm (0 1 gpm) <! 9 I/mm (0 5 gpm)
steam generator (technical gpm)
specifications)
Burst strength
Structural model As design As design Rcinfoiccnicni by TS Plastic limit load With plastic limit load No structural model1
Icak-bcforc-brcak
Margin at nomial operation Design >3 Maigin ,\gamsl lube pull- >1 >1 3, 95% confidence
out
Margin at postulated accident Design >l 4 Against pull-out >l 4 >1 4 1 4, 95% confidence
(SLB)
Experimental background Design Design Pull-out Burst, leak Burst, leak Burst, leak
1
Strongly depends on the inspection hardware and software, which differs in different countncs (USA, France, Belgium )
2
Hot-leg only
TABLE XXV. FTTNESS-FOR-SERVICE GUIDELINES IN EIGHT COUNTRIES
Bases How Implemented Where Used
1 No detectable flaws or leakage • No wall thinning > 20%, no defects over Japan
noise level
2 Flaws limited to a size which is • Use safety factors of 3 for normal USA
calculated not to burst during normal operation and 1 4 and 1 5 for accidents and Canada
operation and accident conditions conservative analysis methods
• Often 40% of wall thickness
3 Flaws limited to a size which is • Use safety factors of 2 7 for normal Germany
calculated not to burst during normal operation and 1 43 for accidents and margins
operation and accident conditions for accuracy and growth
• 50% of wall thickness
4 Flaw limited to a size which is not • Use safety factor of 3 for normal Belgium
expected to burst during normal operation and 1 4 - 1 5 for accidents with Slovenia
operation and accident conditions best estimate analysis, or conservative
analysis methods with no safety factor, use
most conservative result
5 Flaws limited to a size so that there • Use conservative analysis methods for France
is a low probability of tubing burst each degradation mechanism (degradation
during accident conditions specific management) - no explicit safety
factors but aggressive inspections
• Rely on a reliable nitrogen-16 leak
detection system
6 Flaws limited to a size which is not • Use conservative analysis methods Spain
expected to burst during normal supplemented by 100% inspections of
operation and accident conditions affected areas and nght leak rate limits
7 Set defect size based on allowable • Estimate probability of rupture for each Sweden
nsk of rupture during steam line break defect, and require sum for all defects to be
< allowed limit (e g, 1%)
8 No leakage, detectable flaws of any Plug any tube with detectable leakage Russia
size which do not leak are allowed
9 Ensure total leakage for all defects • Estimate leakage for all defects present at Canada
meets dose limits under normal end of interval, make sure total leakage is USA
operating and accident conditions significantly less than applicable site dose Belgium
limits Slovenia
Sweden
110
parameter, obtained from inspection, and experimental results (burst pressure and leak rate
measurements) is derived.
A comparison of the important parameters for each group of repair criteria is given in
Table XXIV.
7.2. TUBING FITNESS-FOR-SERVICE GUIDELINES IN VARIOUS COUNTRIES
The purpose of this section is to review and discuss the steam generator tubing fitness-
for-service requirements in various countries and to describe in a general way how they are
applied. The countries included in this review are Belgium, Canada, France, Germany, Japan,
Russia, Slovenia, Spain, Sweden, Switzerland, and the USA The US Fitness-For-Service
guidelines are discussed first because a number of countries started with the US guidelines
and then modified them. The basis and implementation approach of the Fitness-For-Service
guidelines used in these countries are summarized in Table XXV. Much of the following
material was taken from German et. al., 1994.
7.2.1 Regulatoiy Practices and Fitness-For-Service Guidelines in the USA
Appendix A of Section 10 CFR 50 of the US Code of Federal Regulations requires that

(a) US nuclear power plant owners assure that their reactor coolant pressure boundaries have
an extremely low probability of abnormal leakage, of rapidly propagating failure, and of gross
rupture and (b) the reactor coolant system and associated auxiliary, control, and protection
systems be designed with sufficient margin to assure that the design conditions of the reactor
coolant pressure boundary are not exceeded during normal operation, including anticipated
operational transients. 10 CFR 50 also invokes the ASME Code, including Section XI, which
has specific inspection requirements for steam generator tubing and Section HI, which has
general design guidance appropriate for the analysis of steam generator tubing burst or
rupture. Detailed requirements are contained in each nuclear power plant's Technical
Specifications, which are part of the plant's license from the USNRC, and are often patterned
on the USNRC Standard Technical Specifications.
Article IWB-3521.1 of Section XI of the ASME Code states that the allowable outside
diameter flaws in the tubing in U-tube steam generators shall not exceed 40% of the tubing
wall thickness. Many US nuclear power plants have this criterion in their Technical
Specifications, however, Article FWB-3630 of Section XI does allow alternative criteria to be
used, if approved by the USNRC, and a number of US plants use somewhat higher values in
their Technical Specifications. For example, a 50% wall thickness criterion is used for most
defects at Prairie Island Units 1 and 2, except general wall thinning (Northern States Power
co. 1985). And, a 44% wall thickness criterion is used for all defects at San Onofre Units 2
and 3 (Southern California Edison Co. 1982).
Article F-1341.4 of Appendix F of Section HI of the ASME code limits the applied load
to 0.7 times the plastic instability load, which is determined from either an elastic-plastic
analysis or testing. The plastic instability load is defined in Article NB-3213.26 of Section
111
IQ as the load at which unbounded plastic deformation occurs without an increase in load
This corresponds to a safety factor of 1.0/0.7 = 1.43 for design basis accident loads such as
the loads that would be applied during a main steam line break. If the steam generator tubing
Detailed fitness-for-service guidance is provided to the US nuclear power plant owners

in Regulatory Guide 1.121 (USNRC 1976). However, it should be noted that the USNRC
regulatory guides are not mandatory and the legal requirements applicable to a plant are those
in its Technical Specifications, which are reviewed and approved by the USNRC. Regulatory
Guide 1.121 suggests that three factors be considered when developing a fitness-for-service
limit: "(1) the minimum tube wall thickness needed in order for tubes with defects to sustain
the imposed loadings under normal operating conditions and postulated accident conditions,
(2) an operational allowance for degradation between inspections, and (3) the crack size
permitted to meet the leakage limit allowed per steam generator by the technical specifications
of the licenses" (USNRC 1976).
The minimum acceptable wall thickness is defined in such a way that:
• Tubes with part through-wall degradation should not be stressed beyond the elastic
range during normal operation;
• There is at least a margin of safety of 3 against tube rupture or burst during normal
operation;
• There is at least a margin of safety of 1.43 against tube rupture or burst during
design basis accidents (ASME Code, Section m, Articles NB-3225 and F-1341.4)
• Any increase in the primary-to-secondary leak rate must be gradual enough to allow
corrective actions to be taken prior to tube failure.
The method used to estimate the operational allowance (fraction of the total thickness
to compensate for degradation during the next operating period) should be based on
evaluation of the continuing degradation rate and a consideration of measurement error. A
defect "that reduces the remaining tube wall thickness to less than the sum of the minimum
acceptable tube wall thickness plus the operational degradation allowance is designated as an
unacceptable defect" and a tube with that defect "should be plugged" (USNRC 1976).
Regulatory Guide 1.121 also suggests that conservative analytical models be used to
establish the minimum acceptable tube wall thickness. The wall thickness must meet the
design limits of the ASME Code as discussed above and the stress calculations for defective
tubes must consider all stresses and deformations expected during several design basis
accidents. A summary of the analysis performed must be provided to the USNRC when
applying for alternate fitness-for-service guidelines.
Regulatory Guide 1.121 also suggests that the primary coolant system-to-secondary
coolant system leak rate in the Technical Specifications be adjusted so that the allowable
leakage rate during normal operation is less than "the leakage rate determined theoretically
or experimentally from the largest single permissible longitudinal crack," so that remedial
112
action can be taken if the cracks propagate suddenly. Also the leak rate should be less than
the permissible leak rate based on the site boundary radiation dose. Although not included in
Regulatory Guide 1.121, practice in the USA has been to ensure that the total of all primary
system leaks will not exceed the site dose limits set by 10 CFR 100 during design base
accidents, especially the main steam line break.
7.2.2. Alternative US Fitness-For-Service Guidelines for Outside Diameter IGSCOIGA at

Tube Support Hates
The tubes in the steam generators at the Trojan Nuclear Plant (now shut down) near
Portland, Oregon, experienced considerable outside diameter IGSCC/IGA damage at the tube
support plate locations. Because of the large number of tubes with suspect indications, the
Trojan staff eventually decided to develop alternate fitness-for-service guidelines based on
bobbin coil voltage, to limit the number of tubes requiring repair or plugging (Westinghouse
1991). This involved correlating bobbin coil eddy current voltage with burst pressures. The
experimental work was performed by Westinghouse using pulled tubes from a number of
PWRs as well as model boiler tubes tested at room temperature without tube support plate
reinforcement. The results were then adjusted to the operating temperature of the steam
generator using Alloy 600 temperature-dependent mechanical properties. A lower 95% curve
was established and the voltages corresponding to (a) three times the normal operating
pressure difference across the tubing walls and (b) 1.43 times the pressure drop during a main
steam line break were determined. The expected growth rate in volts during the next operating
cycle was computed based on limited prior experience and subtracted from the burst
correlation results, along with an allowance for measurement uncertainty (analysis, probe
wear, and calibration standards). Because of the uncertainty in the growth rate and the limited
experience with this type of guideline, a limit for the bobbin coil voltage of 1.0 volt was
chosen.
The Trojan alternate fitness-for-service guidelines required an extensive inspection

programme. All the tubes were inspected during the next outage with bobbin coil eddy current
equipment. All indications and all intersections up to the 5th tube support plate were then
inspected with rotating pancake coil eddy current equipment. Care was taken to determine
when a bobbin coil signal was a "possible indication" and a rotating pancake coil signal was
something other than background noise. To provide additional assurance that the Trojan
outside diameter IGSCGTGA defects would be detected, the allowable primary-to-secondary
coolant system leak rate was reduced to 492 litres (130 gallons) per day per steam generator
and 1514 litres (400 gallons) per day for all four steam generators and nitrogen-16 monitors
were installed on the main steam lines.
In parallel with the work at Trojan, EPRI commissioned a committee of US and foreign
experts in steam generator tubing degradation issues to recommend an alternative fitness-for-
service guideline for outside diameter IGSCC/IGA defects (EPRI 1993a, EPRI 1995b). Their
approach is intended to be consistent with the NRC Regulatory Guide 1.121 and the ASME
Section HI philosophy, and is based on experimentally determined voltage limits, as follows:
V=V -
113
Where V is the voltage limit for repair, VSL is a voltage structural limit from correlations of
burst pressure versus bobbin coil measurement, V^ is the measurement error, and V^ is the
voltage growth associated with the expected crack growth during the next operating cycle.
The values for V^ and VSL were taken from an EPRI database, and V^ is based on either
plant specific data or conservative values developed by EPRI. All indications with a bobbin
coil eddy current voltage greater than the calculated limit require repair. The development of
these correlations is discussed in somewhat more detail in the next few paragraphs.
To develop the burst pressure versus bobbin coil eddy current correlation, EPRI compiled
results from pulled tubes from a number of plants and model boiler tubes with diameters of
3/4-inches and 7/8-inches. The eddy current measurements were generally made before the
tubes were removed from the steam generators. The burst tests were performed at room
temperature without tube support plates. Tube support plate reinforcement was not used
because eggcrate and quatrefoil tube supports do not provide coverage around the entire
circumference and drilled hole tube support plates may move during a main steam line break
accident, thereby uncovering the cracks. A curve fit at the lower 95% prediction interval was
then determined and adjusted to 344°C (650°F) using lower bound temperature dependent
mechanical properties. The lower value of the bobbin coil voltage at three times normal
operating pressure or 1.43 times the main steam line break pressure drop was then
determined. This was found to be 4.0V for the 3/4-inch tubing and 4.5V for the 7/8-inch
tubing (for the data available at that time).
To calculate the expected voltage growth during the next operating cycle, data from
consecutive operating cycles at six plants was evaluated. The per cent voltage growth for each
indication and an average plant value was calculated for each plant, then a bounding average
growth rate greater than all the plant average values was computed This value was shown
to be 35%/EFPY.
The measurement uncertainty associated with probe wear and analyst interpretation was
also determined. Probe wear was varied from 0 to 0.5 mm on the centering buttons and
numerous scans made of a four hole calibration standard. The calibration standard was
prepared per ASME Code Section V, Article 8, Appendix n - 860.22. A standard deviation
of 7% was determined. The largest 592 indications from a plant with confirmed outside
diameter IGSCC/IGA at the tube support plates was evaluated by six analysts. The standard
deviation of the voltage readings was found to be 10.3% A combined root-mean squares
deviation for measurement error of 20.5% was then calculated.
Using the EPRI values discussed above and an expected cycle length of 1.3 EFPYs, the
repair limits are 2.4 volts for 3/4-inch tubing and 2.7 volts for 7/8-inch tubing. It should be
noted that use of these limits requires 100% bobbin coil eddy current inspection, and
supplemental rotating pancake coil eddy current inspections to characterize the indications as
outside diameter IGSCC/IGA.
To use the EPRI voltage criteria, the primary-to-secondary coolant system leakage during
various design basis accidents must also be estimated. Therefore, EPRI correlated leak rate
and probability of leakage with bobbin coil voltage by testing pulled tubes and model boiler
tubes at estimated main steam line break pressure differences of 16.1 and 18.3 MPa (2335 and
114
2650 psi). (However, not all the tubes were tested at both pressures, and analytical
adjustments were used.) Using the best fit curves and standard deviations, along with the
crack growth rate and measurement uncertainty distribution, a Monte Carlo analysis can be
performed to calculate an accident leak rate at the end of the next operating cycle. (The leak
rate for a given crack size is the probability of leakage multiplied by the leak rate.) The sum
of the upper 95/95 probability/confidence level values is then used as the conservative upper
bound leak rate and compared to the site boundary limits.
Also, to minimize the probability of rupture, the EPRI guidelines recommend that the
allowable steam generator leak rate be reduced from 1893 to 568 litres per day (500 to 150
The USNRC provided guidance to US utilities who wished to implement the EPRI (or
similar) alternative fitness-for-service guidelines for ODSCC at tube support plates in Generic
Letter 95-05 (USNRC 1995a). Although the USNRC approved the basic approach discussed
above, a number of key parameters were modified and made more restrictive (conservative).
Also, the repair criteria discussed in Generic Letter 95-05 only applies to Westinghouse-
designed steam generators with 19 mm (3/4 inch) and 22 mm (7/8 inch) diameter Alloy 600
tubes and drilled-hole tube support plates and axially oriented ODSCC confined within the
tube-to-tube support plate intersections.
The USNRC voltage repair limits are:
• 22 mm (7/8 inch) diameter tubes with bobbin coil probe indications less than 2.0 volts
may remain in service.
• 19 mm (3/4 inch) diameter tubes with bobbin coil probe indications less than 1.0 volt
may remain in service.
• Tubes with bobbin coil indications greater than the above values but less than an upper
voltage repair limit (calculated using the basic EPRI approach) may remain in service
if a subsequent rotating pancake coil probe inspection does not confirm the indication.
• Tubes with other indications (above the upper limit, or between the lower and upper
limit and confirmed by rotating pancake coil inspection) must be repaired
As with the EPRI fitness-for-service guidelines, the upper voltage repair limit is
determined by first determining the lower 95% prediction boundary for an appropriate set of
room-temperature burst pressure versus bobbin coil voltage data, then reducing this lower
limit to account for the lower 95/95% tolerance bound of the tubing material properties at
343°C (650°F). The structural limit voltage, VSL, is then determined for a free span burst
pressure of 1.4 times the differential pressure calculated for a main steam line break design
basis accident. The structural limit voltage is then reduced to account for flaw growth during
the next operating cycle and voltage measurement uncertainty. The flaw growth allowance
should be based on the voltage growth rates observed at that plant during the last one or two
inspection cycles or 30% per effective full power year, whichever is larger. The voltage
measurement uncertainty should consider probe wear and the variability among data analysts
and should be the 95% cumulative probability value (about 20%).
115
The total leak rate during a main steam line break accident must also be calculated by
(a) determining the frequency distribution of the bobbin coil voltage indications, (b)
determining an end of cycle distribution based on the expected crack growth and estimated
measurement error and (c) use of empirical probability of leak and leak rate versus bobbin
coil voltage indication models. The total leak rate must, of course, be within the licensing
basis. The beginning of cycle bobbin indication frequency distribution must be scaled upward
by a factor of I/POD to account for non-detected cracks, where POD is the probability of
detection of ODSCC flaws and can be assumed to be 0.6. Monte Carlo techniques can then
be used to project the beginning of cycle voltage distribution to the end of the cycle, using
the expected crack growth values and measurement uncertainties discussed above. Once the
projected end of cycle voltage distribution is determined, the leakage is calculated by
multiplying the voltage distribution by (a) an empirical probability of leakage as a function
of voltage value and (b) an empirical leak rate as a function of voltage value. These empirical
models should be developed from appropriate experimental data from either 22 mm or 19 mm
tubing, as applicable.
Implementation of voltage-based repair criteria must include enhanced inspections. The

bobbin coil inspection should include all the hot-leg tube support plate intersections and all
the cold-leg intersections down to the lowest cold-leg tube support plate with known ODSCC.
In addition, 20% of the tubes should be inspected over their full length with a bobbin coil
probe. Rotating pancake coil inspections should be performed for all bobbin coil indications
exceeding 2.0 volts from 22 mm (7/8 inch) diameter tubes or 1.0 volt from 19 mm (3/4 inch)
diameter tubes. Also, rotating pancake coil inspections should be performed at all intersections
with (a) interfering signals from copper deposits, (b) dent signals greater than 5 volts, or (c)
large mixed residuals. Any indications found at such intersections with a rotating pancake coil
should cause the tube to be repaired or plugged. The bobbin coil should be calibrated against
the standard used to develop the voltage-based approach. Probe wear should be controlled.
The data analyst's performance should be consistent with the measurement uncertainties used
Implementation of a voltage-based repair criteria must also include a programme of

steam generator tube removals and testing. Two tubes (at least four intersections) must be
removed from each plant when the voltage-based repair criteria is first implemented. An
additional tube (at least two intersections) must be removed during each outage following 34
effective full power months of operation, or after three refueling outages, whichever is shorter.
The removed tubes must be subjected to leak and burst tests under simulated main steam line
break conditions. The tube intersection areas must also be destructively examined to confirm
that the degradation is axial ODSCC.
Implementation of voltage-based repair criteria must also include reduced leakage limits
(568 I7d or 150 gal per day) and adequate leakage monitoring equipment. Also, tubes with
known leaks must be repaired or plugged
7.2.3. Alternative US Fitness-For-Service Guidelines for PV\5CC in the Roll Transition

Region Proposed by EPRI
Primary water stress corrosion cracking has been found in the roll transition region of
full- and part-depth rolled PWR steam generators worldwide. One of the first alternative
116
fitness-for-service guidelines was the F* criterion which is being used in a number of US
plants. The F* criteria applies to steam generators with partial or full tubesheet depth hard
rolled tubes and allows defects, regardless of size, detected below a certain distance from the
bottom of the roll transition or top of tubesheet, whichever is lower, to remain in service. The
F* distance is established by considering the length of roll expansion needed to resist tube
pull out forces and is typically 38 mm to 50 mm (1.5-2.0 inches), hi other words, the F*
criteria has been applied at locations where there is a very low possibility of steam generator
tube rupture or burst because the defect remains tightly enclosed within the tubesheet.
Recently, EPRI also commissioned a committee of US and foreign experts in steam

generator repair issues to develop alternative fitness-for-service guidelines for tubes with axial
PWSCC above the F* distance (EPRI 1993b). The following equation is used to find the
largest allowable axial crack which can remain in service:
A = a + a-ps
Where A is the allowable crack length, a is a reference crack length from a rupture
correlation, a^ is a correction for tubesheet constraint, a^ is the allowance for crack growth
during the next operating cycle, and a,^ is a measurement uncertainty factor. To develop the
rupture correlation (a versus burst pressure) EPRI compiled results from tests performed on
3/4-inch and 7/8-inch tubing by BELGATOM, Framatome and EOF, Westinghouse, and
CEGB in Great Britain. The data were normalized and a bounding equation determined and
then adjusted to the steam generator operating temperature.
The tubesheet correction factor was developed by BELGATOM as follows (for 3/4-inch
tubing):
0 < a < 4.5 mm a^ = 4.5 mm

4.5 < a < 18 mm a^ = 6.0 - a/3 mm
18 mm< a a^ = 0
These values reflect the tubesheet reinforcement provided relatively short axial cracks
at the roll transition.
The allowance for average crack growth during the next operating cycle (3^) was
determined to be 0.76 mm/EFPY using data from Doel 2 for cracks with beginning-of-cycle
lengths of 3 mm to 11 mm (Doel 2 has an inlet temperature of 330°Q. However, use of plant
specific data is recommended. To determine the measurement uncertainty (a^x), EPRI
compiled results from comparisons of true crack length with crack length as measured by
rotating pancake coil eddy current equipment in France, Belgium, Sweden, Spain, and the
USA The 201 data points provided the following relationship: True crack length equals the
eddy current measured crack length less 0.39 mm with a two sigma distribution of 2.12 mm.
Subtracting 0.39 from 2.12 mm provides an NDE error estimate of 1.73 mm [i.e., the average
true length is 0.39 mm shorter than the measured length but at the 95% confidence level (2
sigma) the true length is 1.73 mm longer than the measured length.]
117
Using safety factors of 3 on the normal pressure drop and 1.43 on the design basis
accident pressure drop, a critical crack length for a 7/8-inch tube of 10.7 mm is calculated.
Use of this criterion required a 100% rotating pancake coil eddy current inspection of all in-
service hot-leg tube expansion zones. It also required a primary-to-secondary coolant system
leakage calculation, similar to the leakage calculation discussed in Section 6.2.2 above.
Also, to minimize the probability of rupture, EPRI recommends that the leak limit during
normal operation be reduced to 568 litres per day (150 gpd) per steam generator.
The USNRC has not approved use of the EPRI proposed fitness-for-service guidelines
for PWSCC in the roll transition region in the US. As discussed below, certain other countries
are using variations of these guidelines.
7.2.4. Other Alternative Fitness-For-Service Guidelines in the USA
Extensive pitting in the Indian Point Unit 3 steam generator caused by a large Hudson
River water excursion into the secondary coolant system in 1981, resulted in alternative
fitness-for-service guidelines at Indian Point Unit 3 during the period 1981 to 1985. Limits
of 65%, 50%, 55% and 63% of the tubing wall thickness were used at various times. These
limits were based on burst testing of pulled tubes and various estimates of next cycle crack
growth and measurement uncertainty. In 1985, Indian Point returned to the ASME 40%
criterion and repaired or plugged all tubes with indications over 40% of the wall thickness.
Extensive circumferential IGSCC/IGA was found in 1991 on the outside surfaces of the
tubes in the three North Anna Unit 1 steam generators. These defects were located in the
tubesheet expansion region and directly above and below the hot-leg tube support plates. All
tubes with significant indications were plugged. However, due to the rapid increase in the
extent of the stress corrosion damage, the plant operator decided that a mid-cycle outage in
10 months was needed.
In an attempt to justify a normal 18 month fuel cycle, the plant operator burst tested
pulled tubes and reevaluated their previous NDE data to develop a conservative crack growth
correlation (over 50% of the 1991 indications were identifiable in 1989). Also, analysis and
testing were performed to determine if fatigue at the defect locations could lead to tube
rupture. Despite these efforts, it was decided that a mid-cycle inspection was necessary. The
plant operator concluded that the results of that inspection showed that the models developed
at the end of the previous operating cycle overestimated the number and size of the tubesheet
expansion zone defects, but underestimated the number and size of the defects near the tube
support plates. The mid-cycle and future inspections of North Arma Unit 1 consisted of 100%
full length bobbin coil eddy current inspections, 100% 8 * 1 probe inspections of the hot-legs,
100% rotating pancake coil inspections near the top of the tubesheet, and follow-up of all
bobbin and 8 > < 1 probe indications with a rotating pancake coil eddy current inspection.
7.2.5. Regulatory Practices and Rtness-For-Service Guidelines in Belgium
The starting point for the Belgian fitness-for-service guidelines were the original US
requirements discussed in Section 7.2.1 above. However, the Belgian regulators consider the
118
40% of tube wall thickness limit in Section XI of the ASME Code too conservative for some
locations and some defect types, and too inflexible. For these reasons, they have revised their
requirements for in-service inspection of steam generator tubes and have defined the
objectives to be met, but assigned the responsibility to the plant operator to meet them. The
revised technical specification
states that the objectives of inspection are to: (1) determine whether tube degradation is
occurring and identify the specific modes involved, (2) assess the rate of defect growth
and compare it with values used in establishing plugging/repairing criteria, and (3)
identify the tubes that require plugging/repairing.
defines the content of the inspection programme, which must include: (1) definition of
inspection techniques and procedures, (2) tubes and zones to be inspected and (3) the
plugging/repairing criteria to be used for each type of degradation.
gives general requirements: (1) requiring inspection methods to be selected such that
they can reliably detect defects of concern, and (2) establishing the minimum sample
size for inspection (3%) and requiring the sample size to be expanded and additional
inspection to be used, if necessary, to achieve the objectives.
Based on the revised inspection requirements, alternate plugging/repairing criteria, i.e.

defect specific fitness-for-service guidelines, have been developed by the plant operator
aiming at (1) ensuring the structural integrity of the tubes, with adequate safety margin, under
normal and during postulated accident conditions and (2) limiting the total primary-to-
secondary leakage during and following an accident to a value consistent with the offsite dose
limit. The controlling accident is considered to be a feedwater/steam line break. With regard
to the safety factors, the Belgian regulators generally use the safety factors on loadings
required by the USNRC Regulatory Guide 1.121 and the ASME Code.
The fitness-for-service guidelines are submitted to the safety authority for approval. They
are reassessed after each inspection to take into account the latest degradation growth rates,
the accuracy of the inspection technique, and any change in the acceptance criteria (e.g.
additional burst test data).
The fitness-for-service guidelines in Belgium are both defect specific and location
specific. For example, one type of fitness-for-service guidelines is for axial PWSCC at the
roll transition at the top of the tubesheet of full depth rolled tubes. Other fitness-for-service
guidelines have been developed for axial IGMGSCC at tube support plate intersections, for
IGA/IGSCC in the sludge pile, and circumferential ODSCC at the roll transition at the top
of the tubesheet. The fitness-for-service guidelines are of two general types, deterministic and
statistical. Deterministic fitness-for-service guidelines are used when the morphology of the
defect is such that reliable sizing is possible using available non-destructive examination
methods, and the size of the defect can be reliably correlated with tube burst data. In these
cases, the measured size is compared to an allowed defect size, which includes margins for
sizing error and growth up to the next inspection, and required safety factors. Statistical
fitness-for-service guidelines are used when accurate defect sizing by non-destructive
examination technique is not possible. In this case, a statistical correlation is developed
119
between a measured non-destructive examination parameter, such as bobbin coil voltage
amplitude, and the burst strength of tubes with defects generally obtained from tubes removed
from service. The lower confidence limit of this correlation, when combined with the required
safety margin, is the maximum permissible value of the non-destructive examination
parameter at the next inspection, i.e. after allowing for growth.
Predicted primary-to-secondary leakage during accidents is calculated on a combined

deterministic-probabilistic basis, taking into account the measured crack size or measured non-
destructive examination parameter at the start of the operating interval, probable crack sizes
or non-destructive examination parameters at the end of the operating interval, and probable
leakage behaviour based on tests of tubes removed from service.
7.2.6. Regulatory Practices and Fitness-For-Service Guidelines in Canada
The regulatory requirements for steam generator tubing fitness-for-service assessments

in Canada are stated in Clause 14 of CAN/CA N285.4, "Periodic Inspection of Nuclear Power
Plant Components" (Canadian Standards Association 1994). In general, the only flaw
indications allowed in unrepaired tubes are where the predicted wall loss does not exceed
40% of the nominal wall thickness prior to the next inspection. However, the most recent
version of this standard allows for indications exceeding the basic 40% criteria, when a
satisfactory fitness-for-service assessment is performed. The fitness-for-service methodology
discussed below is based on recent assessments carried out at Ontario Hydro for Bruce-A Unit
2 and Bruce-B (German et al. 1995).
The fitness-for-service assessment requires demonstration that the incremental risk

associated with continued operation of a steam generator with a known degradation
mechanism is justified, understood and controlled. This has led to the following acceptance
criteria;
1 Demonstrate that the predicted probability of steam generator tube rupture remains
unchanged, thus ensuring that the frequency of the event is unchanged from that
considered in support of the operating license.
2(a) Demonstrate, for all design basis events with possible induced tube failures, that there
are justifiable margins between estimated doses due to consequential tube leakage and
the applicable dose limits.
2(b) Demonstrate that post-accident operating conditions are manageable and procedures
adequate in such a way that overall consequences remain acceptable.
It has been the experience to date that the response to design basis events (criterion
2(a)), has defined the permissible degree of tube degradation in affected CANDU plants
(Grant 1994). Criterion 1 has been demonstrated for the degradation mechanisms experienced
by CANDU units to date, but it may not always be possible to do that for all degradation
mechanisms which might affect the tubes in the future.
120
CANDU plants routinely monitor steam generator leakage with methods capable of
detecting leaks below 10"3 kg/s. Plant operating procedures require shutdown when the leak
rate exceeds 15 kg/h. However, the correlation of leak rate with degradation is usually poor,
because of the dependence of the leak rate on other variables such as applied loads, crack
morphology, crud, etc. Leak monitoring is a useful precaution, but it does not in itself
preclude the existence of large flaws or tube ruptures, it needs to be supplemented by other
actions such as in-service examinations (Grant 1994).
Criterion 2(a) leads to the development of a Maximum Allowable Leak Rate per unit
against which a Total Estimated Consequential Leak Rate due to an event is compared.
The general fitness-for-service assessment methodology consists of the following steps:
Determination of Degradation Mechanism and Root Cause. The first step in the
assessment process determines the degradation mechanism and the root cause of the problem
(more than one degradation mechanism may be affecting the tubes). This leads to two
possible paths of action depending on how widespread the problem is. If the population of
affected tubes is known to be small, i.e. the degradation is not generic, then the affected tubes
are either taken out of service or are allowed to remain in service if the tube flaw indications
are less than the 40% plugging criterion. Actions may also take place to remove the cause of
the problem. For example, if the degradation was caused by debris, there is an effort to
remove it from the steam generator. If the tube degradation is found to affect a large
population of tubes in the steam generators, i.e. the degradation is generic in nature, then the
fitness-for-service assessment continues.
Failure of the Tube(s) under Normal Operating Conditions. For normal operating
conditions, it is necessary to determine the specific degradation mechanism(s); characterize
the tube flaw characteristics and material properties, the loadings and the tube behaviour
under such loadings; and, determine the mode(s) of failure of the degraded tube and the
resulting leak rate. Sources of information for this step include non-destructive and destructive
(tube pulls) examinations and structural testing of tubes containing characterized defects. This
information is required to demonstrate that degradation induced failure of the tube will occur
in a stable controlled manner. This leads to an evaluation of the increase in probability of
boiler tube rupture under normal operating conditions and an evaluation of the adequacy of
the basic Shutdown Leak Rate. Criterion 1 is satisfied if it can be shown that the maximum
predicted probability of boiler tube rupture remains unchanged. This criterion must be
satisfied for the fitness-for-service assessment to continue.
Failure of the Tubes as a Consequence of Design Basis Events, hi order to estimate a

total leak rate for a particular tube degradation mechanism as a consequence of a specific
event, it is necessary to answer the following questions:
(i) How are the tubes likely to fail?
(ii) How many tubes are at risk of failing at the end of the operating cycle,
i.e. just prior to the next inspection?
121
(iii) What remedial measures could be put in place to correct or mitigate the
degradation and what is the impact of these measures on the safety
assessment?
(i) Determine Failure Mode of Tubes
Again, it is necessary to determine the specific degradation mechanism(s); characterize

the tube flaw characteristics and material properties, the loadings and the tube behaviour
under such loading; and, determine the mode(s) of failure of the degraded tube and the
resulting leak rate. This is required to determine the level of degradation beyond which credit
can not be taken for pressure boundary integrity for a particular event (i.e. the tube is at risk
of leaking). This is referred to as the Accident Specific Degradation Threshold Value
(ASDTV). To determine the ASDTV, event specific loadings are considered. These loadings
are obtained from thermal hydraulic analyses of each design basis event and include the
appropriate factors of safety. ASDTV is analytically calculated using flaw models which have
been validated by suitable structural tests.
(ii) Determine Number of Tubes at Risk of Failing by the End of the Operating Cycle
To predict how many tubes are likely to fail by the end of the operating cycle, the future
condition of the tubes must be predicted by determining the present condition, the rate of
change of the degradation and the duration of the operating period to the next inspection. In-
service examination results and probability models provide an estimate of the future condition
of the tubes. In the Bruce-A Unit 2 case the ODSCC was extremely difficult to detect and
required the development of a new eddy current inspection probe, called Cecco-3. This
sensitive probe was a key factor in the success of the overall assessment. The impact of any
remedial action on both the present condition and the rate of change is also included in the
assessment.
In the fitness-for-service assessment, a Maximum Tolerable Flaw Size (MTFS) is also

calculated (based on ductile collapse of flawed tubes) and is used to establish the plugging
criterion which considers the inspection interval, expected growth rate, and inspection
uncertainty. The loadings considered for determining the MIPS are the loadings which
represent bounding loading conditions for the ASME Service Levels A, B, C and D. The
assessment then considers the future condition of the tubes and the calculated threshold value
ASDTV to predict the total number of tubes at risk of leaking by the end of the operating
cycle.
(iii) Remedial Actions
To determine appropriate remedial actions, which could be corrective or preventive in

nature, the degradation mechanism, the root cause and contributing factors must be thoroughly
understood. Further, the impact of all remedial actions must be taken into account in the
safety assessment.
Toted Estimated Consequential Leak Rate. The total Estimated Consequential Leak Rate
is determined by evaluating the product of the total number of tubes at risk at the end of the
122
operating cycle and the total leak rate per tube. Criterion 2 is satisfied if it can be shown, for
all design basis events, that there are justifiable margins between the estimated consequential
doses due to tube leaks and the applicable dose limits and that the overall post accident
consequences remain acceptable.
Fltness-for-Service Assessment. The steam generators in the unit are judged to be fit for
continued service if Criteria 1 and 2 are satisfied. If these relationships are not shown to be
true, then additional measures must be implemented to either further correct the situation (new
plugging limits, internal modifications, cleaning, etc.) or shortening the operating interval
and/or reducing the power levels. The assessment is then repeated.
7.2.7. Regulatory Practices and Rtness-For-Service Guidelines in the Czech Republic
Defective steam generator tubes with 80% or greater wall thickness reduction have been
plugged This value was recommended by the manufacturer, Vitkovice, and has not been
approved by the Czech regulatory body. Additional work to determine the final criterion is
under way (burst testing). Leakage limits were developed and approved by the Czech
regulatory body in 1993.
7.2.8. Regulatory Practices and Rtness-For-Service Guidelines in France
The measures taken by EOF and the French regulatory authority, DSIN, to prevent tube
rupture or burst during normal, off-normal, or accident conditions are based on aggressive
inspection and leak detection programmes supported by defect type and location specific
fitness-for-service criteria (Cochet 1989, Saudan 1992, Lemaire 1993). To find PWSCC in the
roll transition region of steam generators susceptible to PWSCC, the hot leg side roll
transition region of every tube is inspected during each outage using rotating pancake coil
eddy current equipment. To find PWSCC in the small radius U-bends, the U-bend region of
all the tubes in the first row still in service and a sample of the tubes in the second row are
inspected during each outage, using a flexible rotating coil (susceptible steam generators). To
find outside diameter IGSCC/IGA at the tube support plates, the hot leg tube support plate
locations are inspected during every other outage, using bobbin coil eddy current equipment
(susceptible steam generators). Follow-up inspections of affected tubes still in service are
performed during the next outage. Also, 100% of the hot leg tube length in the sludge pile
is inspected using bobbin coil eddy current equipment every other outage. The accuracy of
the examination techniques is assessed by comparing the measurements to the results of
pulled tube destructive examinations (more than 350 to date).
The primary-to-secondary coolant system leak rates in French steam generators at plants
that have experienced tube degradation are measured by y and P spectrometry and nitrogen-
16 monitors, in addition to grab sample measurements.
• When a nitrogen-16 monitor detects a leak greater than a threshold of 20 L/h (lowered
from 72 L/h in 1995), the plant must be shut down immediately.
• When lower leak rates are detected by the nitrogen-16 monitors and are confirmed by
grab sample measurements, plant shut down is required when the following limits are
exceeded.
123
leak rate increase >3 L/h per day, or continuously increasing during 3
days at a leak rate increase >1 L/h per day;
difference between steam generator leak rates >3 L/h;
leak rate >3 L/h (steam generators least susceptible to PWSCC) or >10
L/h (steam generators susceptible to PWSCC).
These stringent total leak rate thresholds and allowable leak rate increases, together with
the maximum allowable background level, are considered important in France as a general
means to improve the surveillance of the tube bundle integrity. Improving the sensitivity of
the monitors so that the occurrence of a potentially new type of growing defect can be
detected earlier, and developing procedures to prevent tube rupture, enhance the safety
margins of the French plants whether or not the flaws which contribute most to the leakage
or the leakage increases are relevant to the leak before risk of break approach, since the
nature and the origin of the flaws cannot be identified during operation. Leak rates models
are presently being developed to better account for the experimentally observed leak rates
from the various kinds of flaws. These models will eventually be used to predict the expected
primary-to-secondary coolant leak rates from the tubing inspection data.
The defect type and location specific fitness-for-service guidelines used in France are
summarized as follows:
Axid PWSCC in the roll transition zone. The axial crack length limit is based on an
analysis of crack growth and tube burst during a main steam line break, which is considered
to be the design basis accident that imposes the highest loads. Correlations of critical crack
length versus burst pressure have been developed from experiment. The analysis uses these
correlations and the maximum tube diameter and minimum wall thickness, adverse mechanical
properties, upper-bound temperatures and pressures, a conservative allowance for crack growth
during the next operating cycle, and margin for NDE error. Safety factors are not applied, hi
developing these correlations, the French specialists concluded that a crack with an end within
the tubesheet where the tube is in contact with the tubesheet can propagate in an unstable
manner in only one direction. The maximum allowable free crack length, which takes into
account accident conditions, is 13 mm for the 900 MWe plants. A temporary criterion of 13
mm has also been adopted for the 1300 MWe plants. A definitive plugging criterion will be
established after completion of certain probabilistic risk studies.
Circumferential PWSCC. Tubes with circumferential PWSCC must be plugged because

(a) the leak before risk of break approach does not apply since the cracks are often not
throughwall until the tube is close to rupture (i.e., the cracks tend to propagate around the
tube first), and (b) the rotating pancake coil eddy current detection limit is only about 50%
of the wall thickness, hi other words, the French experts do not believe that there is much
margin between initial detection of circumferential PWSCC and possible rupture under
extreme accident conditions.
PWSCC in the inner row U-bends. All tubes with indications measured with bobbin coil
and flexible rotating coil eddy current equipment must be plugged. This is because it is almost
124
impossible to pull tubes and develop correlations of actual crack length versus NDE results
and prove that leak before risk of break applies. Preventive plugging of tight U-bend tubes
susceptible to PWSCC was carried out at several French units to improve the availability of
the plants.
Outside diameter IGSCC/IGA at tube support plates. The repair or plugging criteria is
a bobbin coil voltage of 2 volts, which corresponds to approximately 17 volts in the USA.
This rather large limit (as compared with a typical US repair criteria of 1 to 2 volts) is due
to the assumption that the tubing will be supported in those regions by the tube support plates
during various design basis accidents. The 17 volt value was apparently obtained from an
experimental correlation between bobbin coil eddy current voltage and tubing burst strengths
with support plates present, plus a voltage value for the expected crack growth during the next
operating period, plus an allowance for measurement error. A more conservative value (1
volt) is used for defects at the higher tube support plate elevations of some steam generators
(e.g. the model 51A).
Outside diameter IGSCC/IGA in the sludge pile. A bobbin coil eddy current voltage of
500 mV without an axial crack, or 200 mV with an axial crack of 10 mm or greater (detected
by rotating pancake coil), or any ODSCC indication from both bobbin coil and rotating
pancake coil equipment, are the repair or plugging limits for IGSCC/IGA in the sludge pile.
Other indications. All other indications beyond those discussed above, including wear
caused by foreign objects, AVBs, etc., and free-span defects, are judged against the ASME
40% of wall criterion.
7.2.9. Regulatory Practices and Rtaess-For-Service Guidelines in Germany
The repair or plugging criteria in the Federal Republic of Germany are designed to
prevent rupture during normal operation or design basis accident loadings (Azodi et al. 1987).
Defective steam generator tubes are evaluated on a case by case basis. However, wall
degradation of 50% or greater generally results in plugging. This value was obtained from
burst test results and:
• a measurement uncertainty of ±10% for eddy current testing and about ±5% for
ultrasonic examinations
• a factor of safety of 2.7 against rupture during normal operation and 1.43 against rupture
during design basis accident loadings
• an operational allowance for crack growth or additional wastage during the next
operating period of about 3% of the wall thickness (steam generators with phosphate
water chemistry).
7.2.10. Regulatory Practices and Fitness-For-Service Guidelines in Japan
The fitness-for-service guideline issued by the Japanese Ministry of International Trade

and Industry is simply steam generator tubing "flaws are not allowed" (Shizuma 1992). The
125
term "flaw" is interpreted to mean any indication (crack, pit or general wall thinning) greater
than 20% of the nominal wall thickness. Obviously, primary-to-secondary coolant system
leakage is not allowed and a plant with a leak must be immediately shut down upon detection
of the leak. Indications of degradation with a depth less than 20% are considered acceptable
if the eddy current signal shows no change from the previous inspection. Preventive plugging
of tight U-bend tubes susceptible to PWSCC has been performed at two units.
7.2.11. Regulatory Practices and Fltness-For-Service Guidelines in Russia
The fitness-for-service guideline currently used in Russia (and presumably in the rest of
the former Soviet Union) is no steam generator tubing leakage. All tubing with throughwall
cracks which cause detectable primary-to-secondary coolant system leakage are plugged. All
other defect indications are ignored. There has been no sleeving in the WWER steam
generators.
7.2.12. Regulatory Practices and Fltness-For-Service Guidelines in Slovenia
The fitness-for-service guidelines in Slovenia were traditionally based on the 40% tube
wall loss repair criterion. For the power plant located in Krsko, a plant specific value of 45%
was derived and implemented. However, this approach was considered overly conservative
which lead to the implementation of the degradation specific guidelines outlined below. The
defect type and location specific approach is mainly based on extensive inspection (see
section 6.1.7.) and additionally supported by on-line leak detection monitoring (nitrogen-16)
and allowable leak rates of 40 L/h per steam generator.
Axial stress corrosion cracking in the roll transition area The P* and crack length repair
criteria are currently implemented. The P* criterion allows for any defects located at least 38
to 76 mm (depending on the position of the tube) below the top of tubesheet and for axial
defects anywhere inside the tubesheet. The crack length repair criterion is actually based on
the Belgian approach described above and allows for axial cracks in the expansion transitions,
and for both PWSCC and ODSCC in the sludge regions, if the axial cracks are shorter than
6 mm. An additional restraint is that tubes with cracks located more than 7 mm above the
tubesheet have to be repaired if the 45% criterion is violated.
Circumferential stress corrosion cracking in the roll transition area. Any tube with
detected defects of this kind is to be repaired or plugged
PWSCC in the inner row U-bends. Any detected defect triggers repair or plugging of the
tube.
ODSCC at tube support plates. Recently, a conservative version of the EPRI based
voltage methodology has been implemented for ODSCC at the tube support plates. Initially,
100% of the tubes are inspected using a bobbin coil probe. All bobbin coil indications with
a signal amplitude exceeding 1 volt and depth reading exceeding 45% are inspected again
using a multifrequency rotating pancake coil probe. Tubes with defects confirmed by
multifrequency rotating pancake coil probes are then repaired. Probabilistic analyses
addressing events of tube burst and excessive leakage during a steam line break were
performed and are currently being refined to support the implementation of this approach.
126
Other. Tubes with defects exceeding the traditional 45% loss of tube wall thickness are
repaired.
Sleeved tubes. All sleeved tubes and all sleeves are inspected during each outage. A 45%
loss of tube wall thickness criterion is applied for both the intact part of the tube and the load
carrying portion of the sleeve. In practical terms this means repair of all tubes with detected
indications, as no wall depth readings can be obtained from the I-coil and Plus-point eddy
current probes.
7.2.13. Regulatory Practices and Fitness-For-Service Guidelines in Spain
A research programme was launched in Spain to manage steam generator degradation.

Participants included the utilities, manufacturers (ENSA), inspection agency (Tecnatom) and
research centers (Ciemat) [Bollini 1993]. The fundamental objective of the Spanish fitness-for-
service criteria is the same as that of the French criteria, namely, to assure that the critical
crack length under accident conditions is not exceeded during normal operation so that tube
rupture will not occur during a design basis accident. Defect type and location specific fitness-
for-service criteria, along with aggressive inspections of defected steam generators, are used
The Spanish defect type and location specific fitness-for-service guidelines are discussed
below.
Axial PWSCC in the roll transition area Two guidelines are used, the first is the P*
criterion, which allows a tube with axial PWSCC to remain in service if the indication is
below the top of the tubesheet and motion in the vertical direction is controlled by an
essentially nondefective tube. The second guideline is based on the French leak before risk
of break approach which assumes that accurate leak rate measurement during normal
operation will detect crack growth before the crack reaches the critical length. Prirnary-to-
secondary coolant system leak rate correlations as a function of crack size are, of course,
needed for this approach and were based on the French work, modified with Spanish data.
The largest allowable crack length is 8 mm, which is based on a critical crack length of 13
mm (12.6 mm when the rolling is non-standard) less an upper bound crack growth of 4 mm
per fuel cycle and a measurement uncertainty of 1 mm. All tubes with indications equal to
or longer than 8 mm (7.6 mm in the case of non-standard rolling) must be repaired or
plugged. The maximum number of parallel axial cracks in a tube is 20. Tubes with axial
PWSCC in excess of 18 mm above the tubesheet must be repaired or plugged when the defect
is deeper than 40% of the wall thickness.
Circumferential PWSCC in the roll transition area Tubes with circumferential PWSCC
in the roll transition area or above the P* criterion limit must be repaired or plugged. The P*
criterion allows tubes to remain in service if the circumferential indication is located 38 mm
or more (for most of the non-peripheral tubes) below the top of the tubesheet and motion in
the vertical direction is controlled by an essentially non-defective tube.
Outside diameter IGSCC/IGA at the tube support plates. The Spanish utilities have
proposed a fitness-for-service guideline of 78% of the wall thickness for these defects. The
Spanish regulatory agency Consejo de Seguridad Nuclear has not yet approved that criteria
and the Spanish utilities are considering the EPRI bobbin coil voltage criteria discussed above.
127
Other indications. The plugging criterion for fretting damage at the antivibration bar
intersections is 55% of the wall thickness. All indications other than the defect type and
location specific defects discussed above are judged against the ASME 40% of wall thickness
criteria.
The maximum primary to secondary leak rate for steam generators with susceptible
tubing is limited to 5 L/h above a steady leak rate at the beginning of the cycle of 5 L/h or
less. The maximum primary to secondary leak rate for steam generators with Alloy 800M or
thermally treated Alloy 600 tubing is 72 L/h (1728 L/d) during normal operation.
7.2.14. Regulatoiy Practices and Fltaess-For-Service Guidelines in Sweden
The starting point for the Swedish fitness-for-service guidelines are the US requirements,
except that the tubes must be repaired or plugged when the defect indication is greater than
50% of the wall thickness, rather than the 40% specified in the ASME code. However, defect
type and location specific requirements have been developed for axial PWSCC in the
tubesheet region and outside diameter IGSCOIGA at the tube support plates.
The Swedish approach for judging axial PWSCC in the tubesheet region is probabilistic
or risk based in nature (Hedner 1990). The objective is to limit the probability of steam
generator tube burst during a main steam line break to less than 1%, i.e., the sum of all tubes
with an indicated crack length, times the probability of burst for that crack length, must be
less than 0.01.
In equation form;
E u,Px < 0.01
where u^ is the number of cracks of length x and Px is the probability of burst or rupture of
a tube with a crack of length x. The probability of burst includes the expected crack growth
during the next fuel cycle and measurement error and varies as a function of crack length and
distance above the tubesheet. For example, a 12.4 mm crack is calculated to have a
probability of burst of 1%, a 9 mm crack is calculated to have a probability of burst of
0.34%, and a 6 mm crack is calculated to have a probability of burst of 0.0001%. Only the
lengths of cracks above the tubesheet are considered. Tubes with axial cracks below the top
of the tubesheet can remain in service without repair. Tubes with circumferential PWSCC can
remain in service when the cracks are below the P* distance (38 mm below the top of the
tubesheet). Tubes with circumferential PWSCC above the P* distance must be repaired or
plugged
Outside diameter IGSCC/IGA indications at tube support plates with depths up to 70%
of the wall thickness can remain in service provided that the indication is within the length
of the tube support plate and is in the lower five tube support plates. These limits are based
on tube burst testing with a tube support plate present and analysis to determine tube support
plate deflection during design basis accidents. Defects at the upper tube support plates are
allowed when the bobbin coil voltage is less than 1.5V.
128
7.2.15. Fitaess-For-Semce Guidelines in Switzerland
The repair criteria used by the Swiss utility (NOK) is that all tubes with clear indications
within the tubesheet, independent of their depth, will be sleeved and all tubes with indications
outside the tubesheet and greater than 50% of the wall thickness will be plugged
Multifrequency bobbin coil eddy current equipment is used outside the tubesheet region and
rotating pancake coil eddy current equipment is used for supplemental examination of
indications within the tubesheet.
7.3. STATISTICAL METHODS FOR DEGRADATION GROWTH ASSESSMENT
This section summarizes the steps of an overall approach for estimating the rate of
degradation of steam generator tubes. The approach employs a statistical technique and an
empirical model which is consistent with the known degradation processes. A statistical
approach is useful because of the large number of tubes in each steam generator and because
the rate of degradation of steam generator tubes is influenced by a number of materials and
environmental variables. One statistical technique used by some plant operators is the Weibull
probability distribution which is easy to handle mathematically and has been successfully used
to describe the statistics of material failure caused by fatigue and stress corrosion cracking.
An alternate to the Weibull distribution is the log-normal probability distribution. This method
has proven particularly useful for the analysis of laboratory corrosion results (and is, in fact,
suggested by the National Association of Corrosion Engineers for that purpose), and for long-
term projections of degradation in operating steam generators. However, in view of its broader
use in recent analyses, the Weibull distribution will be emphasized here.
The equation for the two-parameter Weibull distribution (Lipson and Sheth 1973) is
F(t) = 1 - expKt/t,)"]
where
F(t) = cumulative fraction of tubes "failed" by a given degradation mechanism
t = time of operation using an appropriate time scale
tr = characteristic time of the Weibull probability distribution (63.2% of a

population has failed by the completion of a period of service equal to the
characteristic time; the value of ^ depends on the environment of the tube at
the failure location)
b = the slope of the distribution when plotted on a Weibull probability graph.
The fraction F(t) in the Weibull equation is the fraction of tubes that are "failed"
according to a particular criterion. Generally, a tube is considered to have failed when it is
removed from service (plugged) or repaired by sleeving because of defects produced by the
degradation mechanism being analyzed. For some purposes, it is useful to use a criterion other
than plugging or sleeving to define the failed condition for analysis purposes. Effective full
129
power years (EFPY) is generally used as a convenient measure of time of operation (total
energy generated divided by the reactor rated power). This measure of time provides an
approximate means of accounting for the effects of changes in operating temperature of the
tubes for different reactor operating conditions. If the reactor has operated for an extended
period of time at substantially reduced power, equivalent full power years should be used.
However, the determination of equivalent full power years requires a value of the activation
energy for the degradation mechanism being analyzed (Shah et al. 1992).
The parameters b and tr in the Weibull equation are adjustable parameters generally
determined by fitting the distribution function to the observed data. The exponent b defines
the slope of the Weibull curve. Its value determines how much scatter there is in times to
failure among a given population. This exponent accounts for the random variations of
properties between different tubes in one steam generator. The characteristic time tr in the
Weibull equation is the basic rate constant of the degradation process. As several of the
degradation mechanisms that affect steam generator tubes are considered to be stress assisted,
thermally activated processes, the parameter tr is primarily a function of temperature, stress,
and chemical environment. For such mechanisms, an Arrhenius equation for the characteristic
time tr is used
t, = A cr"1 exp [Q/R (1/T - 1/TJ]
where
tj = characteristic time appropriate to a specific location

T = temperature for the specific location
o = appropriate stress component for the location
m = constant describing the stress dependence of the degradation mechanism
Q = activation energy of the degradation mechanism
R = gas constant
T0 = temperature for a standard reference condition such as full power condition
A = constant determined from ^ = A (J0~m, where ^ is the characteristic time for
reference condition T0, a0.
Various estimates for the activation energy Q have been derived from laboratory studies
and field experience. For example, the estimate for the activation energy for the PWSCC
mechanism ranges from 39 to 65 kcal/mole, with a best estimate value of 50 kcal/mole
(German et al. 1991, Stein and Mcflree 1986). The stress exponent value (m) is approximately
4 and is briefly discussed in Section 4.1.1. The constant A is a scaling constant determined
by the characteristic time for some standard stress level and reference temperature. The value
of A will change whenever there is a systematic change in the material characteristics and
chemical environment, the average stress level at the location of interest, or other conditions
that may differ from plant to plant.
Figure 24 represents the application of the Weibull model for an assessment of PWSCC
damage to the hot-leg transition region of recirculating steam generator tubes. Tube inspection
data from several plants for PWSCC failure in the hot-leg transition and rolled portion of the
130
0.10
Dashed line
x-- = postpeening
0.01
CO
.A/
I
JO
2
u.
0.001 _ Low tenperature _

mill-annealed tubing
//A .A
Thotieg=616°Fto6190F
(324°C to 326°C)
A'
0.0001
10
Service time (EFPY)
FIG. 24. Weibull analysis ofdataforPWSCC at hot-leg roll transitions and rolled area below
the top of the tubesheet for plants with full depth rolls. (Courtesy of A. P. L. Turner,
Dominion Engineering.)
tube near the top of the tubesheet have been compiled and plotted using a Weibull
distribution. These plants use similar detection technologies and have similar low-temperature
mill-annealed tubing material and primary water chemistry. Therefore all the tubes in all the
plants are within the same PWSCC population. All the tubes are included in the analysis. The
plots illustrate the scatter expected in plant inspection data for PWSCC degradation. The data
for each plant lie approximately on a straight line, except where perturbed by application of
peening as a remedial measure. Even though the intercepts and the slopes (Weibull exponents
b) for each plot vary, the slopes scatter around the bold dashed line drawn for a Weibull
exponent of 3.0. Therefore an exponent of 3.0 can be used to make short extrapolations to
predict the future rate of degradation at a plant where insufficient data are available to
establish a plant-specific slope. However, when looking at a Weibull plot of steam generator
tubing failure data (or eddy current indications of defects) over a longer period of time, the
slope tends to taper off and the rate of cracking is overpredicted when a slope of 3 is used.
This has led some plant operators to use log-normal statistics for projections.
131
Although a number of steam generator experts in the nuclear industry are quite
comfortable with these approaches, some experts at the USNRC and the US national
laboratories question its validity. An Arrhenius equation is an empirical correlation which may
be qualitatively useful, but may not, and in the case of Alloy 600 tubing, has not always
provided accurate life predictions. The failure of the calculated time to correlate with the field
experience may arise from several uncertainties in the input variables. Frequently, the
activation energy is given as 50 kcal/mole, however, the basis for this value is suspect. Some
recent events have suggested that the value is temperature dependent and may be as low as
35 kcal/mole. Additionally, a single value of activation energy may not be valid for both
incubation and crack growth, as is generally assumed when life prediction calculations are
performed. Also, not all of the variables controlling stress corrosion cracking in steam
generator tubes have necessarily been identified, and thus, their omission from the equation
can only lead to erroneous results. This opinion is supported by the consistent failure of
accelerated laboratory corrosion tests to correlate with the Alloy 600 field experience.
7.4. USE OF PROBABILISTIC FRACTURE MECHANICS ANALYSIS TO ASSESS

STEAM GENERATOR AGEING
The development of probabilistic fracture mechanics technologies started in the late

1970s. The first applications in the field of nuclear engineering were concerned with the
reliability of the reactor pressure vessels and primary coolant piping. It has only been in the
last few years that a probabilistic fracture mechanics approach has been applied to the
problem of steam generator reliability, and then only to the question of tubing rupture due to
axial PWSCC at the expansion transition regions (Marko and Cizelj 1992, Pitner et al. 1993).
The basic advantage of a probabilistic fracture mechanics analysis, as compared with a

classical fracture mechanics assessment of design margin, is an ability to address the
uncertainties inherent in the detection and sizing of cracks and in estimating crack initiation
and propagation rates. In other words, all available information (both certain and uncertain)
can be used. The main product of a probabilistic fracture mechanics analysis is a probability
of tubing rupture value. The objectives of the steam generator tubing probabilistic fracture
mechanics analyses are (1) to quantify the influence of ageing management activities such as
tubing inspection and plugging on the probability of steam generator tube rupture, (2) to
compare various ageing management options, and (3) to limit the number of shutdowns
caused by out-of-specification primary-to-secondary leakage (Pitner et al. 1993, Cizelj et al.
1996).
The key items required for a probabilistic fracture mechanics analysis are:
a measured crack length distribution from the most recent tube bundle non-destructive
examination,
the non-detection probability as a function of flaw size and the flaw sizing accuracy,
both from comparisons of pulled tube destructive and non-destructive examination data,
crack initiation rates from comparison of the data from the last two or three sequential
bundle examinations,
132
crack propagation rates from laboratory experiments or again from comparison of the
data from the last two or three sequential bundle examinations,
the probability of detecting a leak before the tube ruptures from laboratory experiments
and the leak detection system sensitivity,
the probability of tube rupture as a function of loading (i.e. the critical crack length) and,
the uncertainties in various dimensions and material properties.
Examples of this type of information can be found in Pitner et al. 1993, Cizelj et al. 1995,
and Cizelj et al. 1996.
Electricite de France has developed a probabilistic fracture mechanics computer code

called COMPROMS which uses this type of information to keep the probability of a tube
rupture in the French steam generators below 10~2 to 10"4 per reactor-year.
The COMPROMtS code calculates three key items:
1) the initial crack size distribution,

2) the change in the crack size distribution with time,
3) the probability of tube failure based on an evaluation of leak before break.
The initial crack size distribution is calculated using a Bayesian approach which
calculates the "a priori" distribution of true sizes from the measured "a posteriori" distribution
and a knowledge of the "likelihood function" (inspection reliability). Probability distributions
of the detection and measurement errors (both systematic and random), which are crack size
dependent, are used to adjust the measured crack size distribution. These probability
distributions are based on comparisons of measured eddy current indications and destructive
metallurgical examination results from about 150 pulled tubes. Also, the initial crack size
distribution is adjusted based on the plugging or repair limits, and the probabilities of
detecting flaws larger than the limit.
The evolution of the initial crack size distribution with time is calculated by considering
the probabilities of new cracks being initiated during the operating period and the expected
propagation of existing cracks. The crack initiation model is based on Weibull statistics
similar to those discussed in Section 7.3 above. The crack propagation model is based on a
correlation between the crack propagation rate and the stress intensity factor which depends,
in part, on the crack size and was developed from laboratory work.
The third key item calculated in the COMPROMIS code is the probability of tube
failure, considering the probability of a detectable leak before break. To obtain the probability
of a detectable leak, a stochastic method is used which considers random factors,
uncertainties, and variabilities in an empirical crack size versus leakage model. A Monte
Carlo simulation then calculates the probability of rupture versus crack size, considering
variations in tube thickness, tube diameter, yield strength, ultimate strength, and tube to tube
support plate contact. Finally, cumulative failure probabilities are calculated as a function of
time.
133
The COMPROMIS code has been successfully used to evaluate the influence of the
inspection strategy (frequency and number of tubes inspected), helium leak testing, the
plugging criterion, and other controllable parameters on the probability of a steam generator
tube rupture. Other probabilistic fracture mechanics analyses have shown that implementation
of Nitrogen-16 monitors leads to about two orders of magnitude lower tube rupture
probabilities and the non-detection probability associated with the existing eddy current
inspections has the largest influence on the tube rupture probability (Cizelj et al. 1996). The
methodology is quite general and can be easily transferred to other combinations of ageing
mechanisms, sites, and inspection technologies and work has recently been directed towards
circumferential cracking at roll transitions and ODSCC at tube support plates.
134
8. STEAM GENERATOR MAINTENANCE: MITIGATION, REPAIR
AND REPLACEMENT
This section discusses mitigation and repair techniques for degradation mechanisms in
tubes, tubesheets, feedwater nozzles and shells. Sections 8.1, 8.2 and 8.3 cover mitigation and
repair techniques for tubes. Table XXVI summarizes countermeasures for tube failures in
PWR steam generators. Most of the measures reduce corrosion by controlling water
chemistry, or by reducing temperature or tensile stresses. Section 8.4 addresses vibration
problems. Sections 8.5 and 8.6 cover mitigation and repair techniques for feedwater nozzles
and shells. Section 8.7 covers replacement measures, such as use of improved tube materials
and modification of the tube supports, which cannot be used in existing steam generators, but
can be incorporated into new designs. The effectiveness of the methods presented in this
section has been mixed, hi general, corrective measures are not as effective as preventive
measures, such as those discussed in Section 5.
8.1. PRIMARY-SIDE MITIGATION TECHNIQUES FOR TUBES
PWSCC may be mitigated by reducing tensile stresses through measures such as

rotopeening or shot peening, stress relieving of the U-bends and controlling the denting
problem. In addition, lowering the operating temperature, protecting susceptible sites (e.g.,
plugging or sleeving, Section 8.3), and reducing the hydrogen concentration in the PWR
coolant helps mitigate PWSCC.
8.1.1 Rotopeening and Shot Peening
Both the shot and rotopeening processes use the impact of a high-velocity small-diameter
mass on the inside surface to produce a layer of cold worked material a few tens of microns
deep. Shot peening uses high-velocity metallic, ceramic, or glass particles. Rotopeening uses
the impact of shots bonded to fabric in a flapper wheel and requires remote tooling in a
radioactive plant. Because there is no post-process nondestructive field inspection technique
to quantify the benefit, the effectiveness of the peening depends entirely on the process
controls. This is a preventive technique, not a repair method for an already cracked tube.
These processes have been used at some units in Europe, Japan, and the USA, but are not
generally used because they are not effective in preventing the growth of existing cracks,
although, they do slow the propagation of old cracks. Also, both rotopeening and shot peening
create outside diameter tensile stresses which could possibly increase the susceptibility of the
tubing to ODSCC. This is recognized in the field process specifications which limit the
outside diameter stress increases to about 4-6 ksi. Also, both processes require remote tooling.
There is very little to differentiate the two methods, although only shot peening is currently
being used.
8.1.2. Stress Relieving
Stress relieving at 705°C for at least 5 minutes also reduces the susceptibility of Alloy
600 to PWSCC, particularly at the U-bends. Laboratory studies indicate that the use of in-situ
stress relief techniques results in at least a factor of 10 increase in the time of PWSCC
initiation. Stress relieving of Alloy 600 tubes in the 650-760°C range may cause sensitization
135
OJ
TABLE XXVI. COUNTERMEASURES FOR TUBE FAILURES IN PWR STEAM GENERATORS

Mechanism Mitigation of damage in existing tubes Improvements in new/replaced steam generators
Primary side stress Rotopeenmg or shot peenmg to residual stresses, stress relieving Alloy 690 tubes with an optimum strength of about
corrosion cracking of the U-bends and control of the denting problem. 380 MPa, little or no residual stresses.
Intergranular stress Control of the alkaline impurities, chlorides, sulphates, and Alloy 690 tubes with an optimum microstructure,
corrosion cracking, carbonates; flush tubesheet crevices; use of hot soak, sludge no tubesheet crevices, improved access for lancing and cleaning,
mtergranular attack lancing, and chemical cleaning; neutralization of crevice increased blowdown capacity, and flow patterns that minimize
alkalinity; addition of bone acid; and full-depth roll expansion of sludge accumulation.
tubes to eliminate crevices.
Pitting Elimination of condenser leakages and ingress of air/oxygen, Titanium or stainless steel condenser tubes, no copper alloys in
chlorides, and sulphates; removal of copper from the feedwater the feed train, and corrosion resistant tube materials (Alloy 690).
train.
Denting Elimination of condenser leakages and ingress of air/oxygen and Strict water chemistry control, stainless steel support structures,
chlorides; use of hot soaks; removal of copper from the feedwater support plates that preclude stagnant water in the annuli, and
tram. titanium condenser tubes
Wastage Use of all-volatile treatment water chemistry; elimination of Flow patterns that minimize hide-out and chemical
hideout chemical concentrations; use of sludge lancing, chemical concentrations and sludge formation; improved access for
cleaning, hot soaks, and hot blowdowns and flushing; preclusion cleaning; increased blow-down capacity.
of resin ingress.
High cycle fatigue and Additional antivtbration bars (AVBs) and insertion of the AVBs
fretting in RSGs deeper into the bundle; minimum tube-to-AVB clearances and
wear matching of the AVBs to the tubes.
Erosion-corrosion Control of the chemistry and entrained solid particle content of the
and corrosion fatigue in secondary side coolant
once-through steam generators
(formation of chromium depleted regions near grain boundaries) and susceptibility to
secondary-side IGSCC/IGA; however, this may not be a concern for Alloy 600 material with
intergranular carbides and low solid-solution carbon content. Thermally treated Alloy 600 and
690 may also be stress-relieved without causing sensitization.
8.13. Reducing the Hot-Leg Side Temperature
Corrosion is a thermally activated process and is, therefore, strongly affected by the
tubing temperatures. The primary coolant entering the hot-leg sides of the RSG, or the top
of the once-through steam generators, is about 25°C to 40°C (45-72°F) hotter than the coolant
leaving the cold leg side of the RSG or the bottom of the OTSGs. Therefore, the hot-leg side
of the tubes in the RSG have usually experienced corrosion degradation much earlier than the
cold leg sides. Similarly, the upper regions of the tubes in the OTSGs have experienced
corrosion degradation, while the lower regions have been relatively immune from these
problems. The primary coolant hot-leg temperatures at full power vary from about 315°C to
327°C (599-621°F) at most PWR plants; initiation of stress corrosion cracking has usually
been later at plants with lower hot-leg temperatures. A few plants have operated with hot-leg
temperatures at the steam generator inlet as high as 330°C (626°F) and have experienced
relatively early and extensive primary and secondary side corrosion. (However, the outside
diameter stress corrosion cracking may have been influenced by factors other than
temperature.) Conversely, the CANDU plants have operated with a hot-leg temperature of
about 308°C (586°F) and have not experienced PWSCC to date.
Reducing the temperature of the tubing on the hot-leg side by about 10°C or more is
believed to slow down, though not preclude, various thermally-activated damage mechanisms
on both the primary and the secondary side. This is a temporary mitigation technique that can
increase the time between the steam generator outages required for inspection. Plant
availability is increased, but this benefit may be offset by reduced power during operation.
However, it is possible to reduce the hot-leg temperature by up to 10°C without lowering the
power at some plants because of the turbine designs.
8.2. SECONDARY-SIDE MITIGATION TECHNIQUES FOR TUBES
The actions which will slow down or prevent outside diameter stress corrosion cracking
(IGSCC and IGA), pitting, and denting of the steam generator tubing are primarily those
discussed in Section 5 above associated with control of the secondary coolant system water
chemistry (chemical additives, actions to limit impurity ingress, and steam generator
cleaning). Reducing the hot-leg temperatures to values at or below about 300°C (575°F) will
significantly slow down thermal-activated damage mechanisms such as IGSCC and IGA, but
not to the same extent as PWSCC (Richards 1985, EPRI 1985). Another secondary side
mitigation technique used on many of the older Westinghouse-type steam generators with
tube-to-tubesheet crevices was full-depth or almost full-depth expansion (usually hydraulic
expansion) of the tubes.
137
8.3. TUBE REPAIR
Damage and failure of steam generator tubes occurs with greater frequency than with
other components, partly because the tubes are exposed to both the primary and secondary
coolant and more than 50% of the pressure boundary is associated with the tubing. For this
reason much of the research and actual repairing in steam generators is associated with the
tubing.
8.3.1. Plugging
Plugging was the only countermeasure available for PWR steam generator tubes with
unacceptable flaws until the early 1980s. Denting has caused several thousand tubes to be
plugged in some plants. Even now, plugging is often done for unacceptable degradation above
the tubesheet region because the current sleeving techniques are difficult or expensive to
implement high up in a steam generator. More than 103 000 plugs are currently installed
worldwide in PWR and CANDU plants (EPRI 1994). Plugs have typically been made from
bar stock of Alloy 600 material; however, most currently installed plugs are made of Alloy
690 material. Commonly used techniques to plug a tube include welding, explosive forming,
and mechanical or rolled installation. A typical mechanical plug in its unexpanded and
expanded forms is shown in Fig. 25.
A plugged tube may continue to be susceptible to stress corrosion cracking, fatigue, and
fretting damage, and finally sever. However, the temperature of the coolant in a plugged tube
is about 40°C (70°F) less than an unplugged tube on the hot-leg side. This will greatly reduce
the PWSCC rates. A severed tube may experience large amplitude vibrations because of
fluid-elastic instability and then damage neighbouring tubes. To prevent this, plugged tubes
can be stiffened by inserting stabilizers, for example, solid rod segments that can be threaded
to each other and to the plug.
..Plug sheU
•Approximate
location of
North Anna-2
plug failure
Tubesheet
'Expander
Unexpanded plug Expanded plug
FIG. 25. Sketches of unexpanded and expanded mechanical plugs (Westinghouse 1989).
Reprinted with permission.
138
Certain types of plugs are also susceptible to PWSCC-type degradation, as described
below:
1. Explosive plugs of Westinghouse design installed in the 1970s have experienced

PWSCC, as evidenced by several reports of leaking plugs. These plugs were made
of mill-annealed Alloy 600. Explosive plugs have leaked in at least three plants
because of large plastic strains and unfavourable residual stresses at plug corners.
The cracking seen in at least one plant was circumferential in orientation and
occurred at the top transition of the explosive expansion, i.e., in a pressure
boundary region.
2. PWSCC has recently been reported as occurring in Babcock & Wilcox

mechanically rolled plugs installed in recirculating and once-through steam
generators, and has been identified in different thermally treated Alloy 600 tube
plug heats (USNRC 1989b). This PWSCC has occurred in the form of
circumferential cracks located in the transition below the roll expansion, i.e., at the
"heel" location, which is not part of the pressure boundary, so that plug integrity
is not affected (see Fig. 26). To a lesser extent, degradation in the form of axial
cracks in the transition above the roll expansion, i.e., in the "toe" location, which
is part of the pressure boundary, has also been identified. The degradation has been
detected by rotating pancake coil inspection. Cracked plugs have only been seen
on the hot-leg sides to date.
3. Mechanical plugs of Westinghouse design have experienced PWSCC in the

expanded area. These plugs have been installed in large numbers since about 1980.
The most significant occurrence of PWSCC in this type of tube plug occurred in
February 1989 at North Anna Unit 1 (USNRC 1989a, USNRC 1990b). The plug
involved was made with thermally treated Alloy 600 material, hi this incident,
circumferential PWSCC occurred nearly throughwall all around the circumference
Roll
transition - Toe
PWSCC
_„—- crack
Roll — ~~ *
,*•
location
^
Plug head
r ^
FIG. 26. PWSCC cracks in rolled plug supplied by Babcock & Wilcox (USNRC 1989b).
139
of a plug, as shown in Fig. 25 (Westinghouse 1989). The remaining ligament broke
during a plant transient and allowed the top part of the plug to be propelled up the
tube until it hit the U-bend, which it penetrated, causing a significant
primary-to-secondary-side leak. The adjacent tube was deformed, but not
penetrated, by the impact of the plug top.
It was originally estimated that mechanical plugs potentially susceptible to PWSCC

had been installed in about 7000 tubes in Westinghouse steam generators in
approximately 20 US PWR plants. However, plugs made with what were thought
to be relatively non-susceptible heats have recently experienced significant
circumferential cracking at only about 20% of the estimated lifetime. The USNRC
now requires US utilities to implement a programme of plug removals, inspections,
and repairs for all Westinghouse mechanical plugs fabricated from thermally treated
Alloy 600. Also, installation of Westinghouse mechanical plugs fabricated from
Alloy 600 material should be discontinued (USNRC 1991).
More than 3000 steam generator tubes in France have also been plugged with Alloy
600 mechanical plugs. Some of the plugged tubes have experienced a "boiler
effect" due to throughwall cracks in the plugs. Removal of these plugs from the hot
leg tube ends is now scheduled
4. Welded plugs manufactured by Combustion Engineering have also degraded and

resulted in leakage in service. Cracking in these Alloy 600 plugs has occurred in
the welded region and is believed to be caused by PWSCC degradation. Plug
defects are repaired by rewelding or replacement with a new welded plug.
Improvements in weld materials and procedures are expected to preclude future
degradation in the new welded plugs. Similar problems have also occurred in
Europe.
Despite the plugging of a relatively large number of tubes, a steam generator may still
generate the rated capacity of electricity because it normally starts operation with a significant
margin of available capacity. However, continued plugging after the margin is exhausted can
significantly reduce plant capacity (Richards 1985). Also, the plugging of a very large number
of tubes can impact the thermohydraulics of a steam generator and result in safety problems.
Before this occurs, the potential economic consequences (cost of anticipated repairs and cost
of lost capacity from plugged tubes) should necessitate extensive sleeving or steam generator
replacement. Steam generators lose their capacity margin when about 5% to 20% of their
tubes are plugged.
Sleeves can be installed into previously plugged steam generator tubes to restore plant
capacity if the plugs are successfully removed. Welded plugs are typically removed by electric
discharge machining (EDM) or drilling or milling techniques. Mechanical plugs can generally
be removed (pulled) with a hydraulic device.
83.2. Sleeving
Sleeves placed inside PWR steam generator tubes bridge over the damaged tube regions.
They are designed to take the full loads that the original tubing was designed to take, i.e., the
140
sleeve replaces the tube as a structural element from its top joint to its lower joint. Most of
the currently available sleeving techniques are designed to cover the inside surfaces of PWR
steam generator tubes in the region from the bottom of the tubesheet to slightly above the
sludge piles. Another location for sleeving is at tube support plate intersections, where sleeves
are used to repair IGA/IGSCC occurring at the tube support plate-to-tube crevice. Sleeves at
tube support plates have only been used on a very limited basis in the USA [originally at
Palisades in the mid-1970s and recently at one or two other plants on a trial basis], but they
are now being extensively used in some Japanese and other plants. Sleeves can be installed
into previously plugged steam generator tubes to restore plant capacity if the plugs are
successfully removed. The sleeve is normally made of a material having better corrosion
resistance than the original tube material, such as thermally treated Alloy 600 or Alloy 690.
Sleeve designs vary primarily in the joint between the sleeve and the tube. The sleeve
top joints can be either the leak limiting type or the leaktight type. The leak limiting type
such as a hydraulically expanded joint and a hybrid expanded joint have been widely used,
and are continuing to be used in relatively low-temperature plants (typical hot-leg temperature
less than 315°Q. However, recently hybrid expanded joint sleeves have been installed in a
high-temperature plant (typical hot-leg temperature in the range of 325°C to 330°C). The
sleeves are generally 760 mm to 914 mm (30-36 in.) long. The hybrid expansion is
constructed by first making an approximately 100 mm- (4 in.-) long hydraulically expanded
section, and then making a shorter 50 mm- (2 in.-) long mechanical-roll (hardroll) expanded
section centered within the hydraulically expanded section as shown in Fig. 27, making a
leaktight or nearly leaktight mechanical joint. Residual and operating stresses in the lower
hardroll transition, shown in Fig. 28, are likely to be the highest in the entire hybrid expanded
joint. The upper joint of the sleeve is provided with an overlengih of 50-112 mm (2-4.5 in.),
which represents a length of unexpanded sleeve above the upper expanded section that should
limit lateral displacement of the tube in the event of a circumferential break of the parent tube
in the upper joint region. Note also that the upper hydraulic transition of the upper joint of
the parent tube represents the primary pressure boundary. Also, the inside surface of the
parent tube along the sleeve overlength is deeply creviced and is in contact with the primary
coolant. Also, the gap between the section of the parent tube which is no longer the primary
pressure boundary and the sleeve is a deeply creviced geometry, and if a defect in the parent
tube grows through the wall, the crevice will be filled with stagnant secondary water that may
promote IGSCC or IGA of both the parent tube and sleeve. The reasons for using a hybrid
expansion are to (a) make the main transition between the unexpanded and expanded areas
by a hydraulic method, which results in lower residual stresses with less potential for stress
corrosion cracking, and (b) include a mechanically rolled area, which results in a tighter joint
with lower leakage than achievable using hydraulic expansion alone.
Sleeves currently being installed in higher temperature plants with PWSCC use leak tight
joints which are made by brazing or welding after the sleeve is expanded to be in intimate
contact with the tube as shown in Fig. 29. Explosively welded joints are expanded and welded
in one operation. Laser welding of the sleeving is being used in the USA, Europe and Japan,
in part, to reduce radiation exposure. The sleeves are typically 900 mm (36 in.) long or longer
to ensure that the top joint is well above the tubesheet and sludge pile areas and there is
sufficient overlength above the top joint. The weld regions must be prepared before sleeving
to remove any oxidation or corrosion layers, thereby ensuring a proper metallurgical bond.
141
Upper Hydraulic Upper Hardroll
Expansion
Parent Tube •
•Sleeve
Lower Hydraulic
Expansion
Cladding
FIG. 27. Typical sleeve installation with a hybrid expanded joint (Westinghouse 1994).
Copyright Westinghouse Electric Corporation; reprinted with permission.
Welded joints in free span areas of high-temperature plants (typical hot-leg temperature of
about 327°Q with PWSCC susceptible tubing are generally stress relieved with a short-time
high-temperature heat treatment, in order to minimize the likelihood of PWSCC.
Westinghouse installed approximately 30 000 laser welded sleeves in the Doel Unit 4 and
Maine Yankee steam generators during 1994 and 1995.
Explosively expanded sleeve joints have also been used to affect a leaktight repair for
degradation both at the top of the tubesheet and at tube support plate elevations; these were
referred to as "kinetic" sleeves, hi the period from 1991 through 1992, more than 2500 such
sleeve repairs were made by Babcock & Wilcox in steam generators at Duke Power's
McGuire Unit 1 and Unit 2 plants, and at Trojan. After three separate forced outages due to
primary-to-secondary leakage, all of the kinetically sleeved tubes were removed from service
by plugging. The cracking that occurred was circumferentially oriented PWSCC of the parent
tube just above the upper explosive joint. Failures occurred both in joints that were not stress
142
Upper Hydraulic
Transition
Upper Hardroll
Transition
Lower Hardroll
Transition
Lower Hydraulic
Transition
FIG. 28. Hybrid expansion joint configuration (Westinghouse 1994). Copyright Westinghouse
Electric Corporation; reprinted with permission.
relieved (due to high local as-fabricated residual stresses) and joints that received a high
temperature post-installation stress relief. The latter failures are probably due to the high
stresses that can develop during thermal stress relief of a tube that is not free to expand
axially due to "lock-up" at the tube support plates (e.g., restraint due to localized denting or
corrosion product buildup).
Another approach to the repair of roll transition region stress corrosion defects in part-
depth rolled tubes is the design used on an experimental basis at Doel Unit 2. This approach
uses the thin Alloy 600 minisleeve shown in Fig. 30; this sleeve is about 40 mm (1.5 in.)
long, explosively expanded and welded over the cracked tube. A portion of the tube that was
not previously expanded is now expanded against the tubesheet, thereby providing the load
carrying capability, hi other words, the transition region between the deformed and
undeformed portions of the tube is now in a new defect-free location, with probably less
residual stresses than the original hard-rolled joint. The sleeve is so thin that the inside
dimension of the tube is practically not altered, allowing full flow of reactor coolant. This
design does not preclude the future use of a longer sleeve with other designs and processes
if new defects are found later.
143
Sleeve
Expansion
Tubesheet
Weld Cladding
FIG. 29. Welded sleeve design (Combustion Engineering) (EPRI1985). Copyright Electric
Power Research Institute; reprinted with permission (modified).
Steam generator ^_. Steam generate

^ tube
tube
Tubesheet
J* ''defect
Repaired tube J
K area I
Roll expansion Mini sleeve
a). Illustration of a local defect b). Defect area after installation

in the roll transition of the tube. of minisleeve.
FIG. 30. The minisleeve design (Babcock & Wilcox) using explosive welding (Gorman and
Mundis 1983). Copyright Electric Power Research Institute; reprinted with permission.
144
Service experience with the almost 100 000 sleeves installed worldwide has generally
been good. However, some parent tubes at at least eight plants have experienced degradation
and were subsequently plugged and removed from service. Trie cracking is mainly caused by
stress corrosion in which the residual stresses introduced by the sleeving process play a major
role. The experience with the kinetic sleeves was mentioned above, examples of other
problems are discussed below.
Use of minisleeves at Doel Unit 2 was discontinued because shortly after their
installation, PWSCC occurred in five expanded tubes at the tops and bottoms of the sleeves.
All tubes with this type of sleeve were plugged. At two other plants, four parent tubes
experienced throughwall cracking at the welded sleeve joints and the sleeves collapsed. At
another plant, the defective welds in two welded sleeve joints has resulted in leakage of
primary coolant.
In 1993, over 1700 tubes in three steam generators at the Doel 4 plant were repaired
with hybrid joint sleeves. During the next refueling cycle, two of the sleeved tubes leaked.
One of the leaking tubes was pulled and examined in the laboratory. The examinations
showed that the leakage occurred at the upper hydraulic transition of the upper joint. The leak
resulted from a 180 degree, throughwall circumferential crack which initiated at the inside
diameter (PWSCQ. Removal and examination of another sleeved tube also showed a 90%
to 100% throughwall circumferential crack extending 30 degrees around the parent tube. The
utility has plugged or laser-weld repaired all the tubes in which hybrid joint sleeves were
installed in 1993.
Out of 75 indications detected in the upper region of the sleeves with hybrid expanded
joints at the Kewaunee plant, 74 were circumferential and one axial. Most of the
circumferential indications were detected at the lower hardroll transition of the joint. These
indications were most likely initiated by ODSCC. Several indications were detected at the
lower hydraulic transition. One circumferential indication was detected at the upper hydraulic
transition, which, as mentioned before, constitutes a primary pressure boundary. The axial
indication was located in the hardroll region and most likely initiated by ODSCC. No sleeved
tube from the Kewaunee plant has been pulled and examined in the laboratory. So the
degradation mechanisms responsible for detected indications have not been identified with
certainty.
Inspection of the parent tube through the intervening sleeve is difficult. This is
particularly true for detection of circumferential cracks at upper transition regions. Special
eddy current probes are being developed for this inspection.
Estimates of the service performance for sleeve repairs of degraded tubes are usually
based on accelerated corrosion tests of mockups prepared to replicate as nearly as possible
the conditions that exist in the field. Recently, this has included testing under conditions of
axial restraint, to stimulate far-field stresses that develop due to lock-up at support plates.
Environments that have been used for these tests include high temperature (400°Q dense
steam and mild caustic (NaOH) solutions. In order to benchmark the results of these tests -
i.e., to provide some basis for interpreting the results in terms of performance in the steam
generator - components whose corrosion resistance in service is known are included in the
test autoclaves. These latter mockups are typically mechanical roll expansions.
145
The anticipated performance of a sleeve (lifetime) depends on the nature of the sleeve
repair (mechanical without stress relief versus a fusion with or without thermal stress relief,
etc.), the location of the degradation, whether the degradation is PWSCC or ODSCC, the
resistance of the parent tube to stress corrosion, the extent of the restraint at nearby tube
support plates, operating temperature, and possibly other factors. Therefore, the lifetime of
a sleeved tube could be as little as two cycles when the parent tube has a low resistance to
stress corrosion cracking, the joints are mechanical and not stress relieved, and the steam
generator is operating at high inlet temperatures (e.g. 330°Q. Or, the lifetime of a sleeved
tube could be as long as 20 years for thermally stress relieved laser welded sleeves in a low
temperature steam generator (e.g. 315°Q.
83 J. Nickel Plating
A nickel plating technique has been developed by Framatome and Belgatom to repair
PWSCC cracks in PWR steam generator tubes. The nickel plating consists of electrolytically
cleaning the damaged surface and then depositing up to about 200 microns of pure nickel on
the damaged surface. The nickel deposit on the damaged tube wall bridges the through-wall
cracks and stops leakage of primary coolant into the secondary system. In addition, the nickel
deposit prevents contact between the primary coolant and the damaged tube wall, arresting
crack propagation and stopping the initiation of new cracks. Nickel plating has been qualified
for steam generator tubes and has been applied to about 1100 tubes in Belgium and Sweden
in the last eight years. All these tubes, except the first few, are still in service, whereas
unplatted tubes are degrading.
Nickel plating has several advantages over sleeving. It generates very low residual
stresses and does not require a subsequent heat treatment, and it can be applied anywhere in
the straight part of the tube. It also allows later access to areas above the plated section for
repair of any damage, whereas sleeving does not. Nickel plating is also a reversible process
because, if needed, the plating can be stripped off chemically without damaging the tube.
Nickel plating has one major drawback, the thin plating does not provide a structurally
acceptable pressure boundary. Therefore, the parent tube has to continue to carry most of the
pressure loadings and parent tubes with large through-wall cracks may not have adequate
strength. With respect to in-service inspection of the plated region, nickel is magnetic and the
nickel layer creates a barrier to the small magnetic field introduced by conventional
eddy-current coils. Therefore, these methods cannot be used to inspect a nickel-plated region.
However, a new ultrasonic inspection method capable of detecting axial and circumferential
cracks has been developed to overcome this problem. Pulsed magnetic saturation eddy-current
techniques may also be used for inspecting nickel-plated tubes [Dodd 1988].
8.4. VIBRATION CONTROL
8.4.1. Preheater Repair
The wear/fretting problem in the Westinghouse type D2/D3 RSGs was addressed by
redistributing feedwater flow between the primary and auxiliary feedwater inlets to reduce the
flow into the preheater through the primary inlet, and by incorporating a preheater manifold
146
to reduce crossflow vibration. Turbulence and peak flow velocities were reduced.
Westinghouse model D4/D5/E RSGs were modified by performing an expansion of the tubes
within the tube baffle plates at certain preheater locations, in effect changing the tube natural
frequency. In addition, a split of the feedwater flow between primary and auxiliary inlets was
also implemented on the D4/D5 RSGs. The inspection results to date indicate that these
modifications have been effective at reducing preheater wear/fretting to a level of minor
concern [Hofmann et al. 1986, Sudduth 1986]. The Krsko plant has operated for about 10
years without any problems in the preheater sections of their Westinghouse Model D-4 steam
generators, after expansion of some of the tubes and split feedwater was introduced.
8.4.2. Antivibration Bar Replacement
Fretting problems occurred in a number of the earlier RSGs and resulted in additional
or longer antivibration bar installations. For example, additional antivibration bars were
installed in the Beznau Unit 1 steam generators in Switzerland. Antivibration bars have been
replaced in Japan, Spain, the USA, and elsewhere.
8.5. MITIGATION OF THE THERMAL FATIGUE OF THE FEEDWATER NOZZLES

AND PIPING
Several modifications in the operation of the auxiliary feedwater systems have been
made to minimize or prevent thermal fatigue of the feedwater piping and nozzles. The
modifications include:
• Use of auxiliary feedwater with a steady flow instead of fluctuating or intermittent

flow.
• Shorter hot standby and low-power operating periods during which auxiliary
feedwater has to be used.
• Use of a nitrogen blanket on the condensate storage tank, which feeds the auxiliary
feedwater system. This prevents ingress of oxygen to the auxiliary feedwater, and
thus reduces corrosion fatigue.
• Use of heated auxiliary feedwater to reduce corrosion fatigue.
• Use of temperature monitoring to control the auxiliary feedwater injection and

thermal loadings.
Several modifications in the design of the feedwater system have also been made in the
existing or new steam generators minimize or prevent thermal fatigue damage to the
feedwater piping and nozzles. Some of these modifications have been also employed for
147
repairing the thermal fatigue damage as discussed in the next section. The modifications
include:
• Replacing the sharp counterbore with a blend radius to reduce stress concentrations.
• Installing a separate nozzle for injecting the auxiliary feedwater directly into the
steam generator.
• Use of a spraying device located in the feedwater piping upstream of the nozzle to
mix the cold auxiliary feedwater with the hot water in the pipe.
• Installing a long thermal sleeve to protect the feedwater piping from thermal
stresses and fatigue damage induced by thermal stratification.
• Welding the thermal sleeve to the feedwater nozzle to reduce stratification in the
annulus region between nozzle and thermal sleeve.
• Use of a destratification loop in the feedwater piping (either just inside or just
outside the steam generator shell).
8.6. REPAIR OF)WATER NOZZLES AND PIPING
At one PWR plant in the USA, the feedwater nozzle bore region, blend radius, steam
generator shell inside surface beneath the nozzle, and feedring support bracket welds
experienced thermal fatigue cracking. All cracks and indications were removed by grinding
Existing Feedwofer Nozzle
• Existing Feedwater Piping
(to be determined irrthe field)

• Existing Feedwater
Nozzle
FIG. 31. Schematic diagram of tuning fork repair used infeedwater nozzles (From Cofie et al.
1994). Copyright American Society of Mechanical Engineers; reprinted with permission.
148
and were repaired by welding to the design configuration. Cracks were also found on the
inside surface of the bore of the inspection ports, which were ground and weld repaired with
the Westinghouse half-bead technique [Stoller 1990].
The feedwater piping-to-steam generator nozzle connections have been particularly

susceptible to thermal fatigue damage, hi the case of D.C. Cook, piping was repaired as
shown in Fig. 12b. The backing strip was removed in the redesign, and the sharp discontinuity
where the crack initiated was replaced with a blend radius. Also, a long thermal sleeve was
installed to protect the counterbore region and a gamma plug was installed for in-service
inspection of the thermal sleeve.
At Diablo Canyon Unit 1, erosion-corrosion of the outside surface of the carbon steel
thermal sleeve caused thinning of the sleeve [NRC Information Notice 92-21, 1992]. This
damage was due to feedwater leakage through the feedwater nozzle-thermal sleeve joint and
the fact that the chromium content of the thermal sleeve material was less than 0.1 weight per
cent (Section 45.4). Thermal sleeves are being repaired with a new tuning fork design, shown
schematically in Fig. 31. The tuning fork is an integrated single piece, welded to the existing
piping. It fits over the old thermal sleeve, preventing bypass leakage between the existing
nozzle and thermal sleeve [Burns 1994]. There is a region of stagnated water in the gap
between sleeve and nozzle, but no flow [Cofie et al. 1994]. Several plants have also replaced
the affected portion of the piping adjacent to the feedwater nozzle.
8.7. STEAM GENERATOR REPLACEMENT
The loss of power attributed to plugging of the steam generator tubes or the effort
involved in sleeving a large number of tubes may not be acceptable, hi such a situation, the
following three alternatives are available: replacing the entire steam generator, replacing the
lower assembly of the steam generator, or retubing the steam generator using the existing
tubesheet and shell structures.
Replacing the entire steam generator involves cutting the piping (primary and secondary),
forcing the generators from their supports, and removing them through existing equipment
hatches in the containments. Where existing equipment hatches cannot be used, considerable
work may be involved in cutting temporary transfer openings in prestressed concrete
containments. The replacement of PWR steam generators also introduces difficulties in fitup
and welding not encountered in previous nuclear plant maintenance or repair projects and
requires considerable decontamination work. On the positive side, because utilities are now
exchanging steam generator replacement experience, the outage times caused by replacement
have been significantly reduced. For example, the Gravelines 1 steam generator replacement
was completed in 40 days.
Replacement of the lower assembly of a recirculating steam generator involves cutting

the steam generator at the transition cone and removing the lower assembly, including the
tubesheet forging, tubes, etc., and then replacing it with a new lower assembly. This
procedure has been chosen for Turkey Point and North Anna Unit 1 because it was difficult
to cut the reactor coolant piping because of access problems. Instead, a channel head cut was
used; however, cladding the inside surface of this weld proved to be more difficult than
anticipated.
149
In-place steam-generator retubing involves cutting and removing the tubes and
tube-support structures and replacing the steam drying and separation equipment. Retubing
has not been used at any utility because it requires a two- to three-year outage period, the cost
of the lost power during the outage is significantly higher than the cost of the other
repair/replacement options, the site work involves as much fabrication as in manufacturing
a new steam generator, and the process controls may not be as effective in the field as in the
shop. This option also involves a significantly large radiation exposure to the personnel.
8.7.1. PWR Replacement Steam Generator Designs
Steam generator replacement is expected to result in a longer steam generator life than
repair because design and materials improvements (some of which are listed in Tables I
through IV) can be implemented, and the impact of prior operating history is removed. Some
of the improvements that are being incorporated into new PWR steam generators are
discussed in this section.
Design and manufacturing modifications to reduce residual and vibrational stresses and
corrosion sites are being used. The residual stresses at the expansion transition region and at
Flow Lines
Sludge
Flow
Flow lines Distribution
Baffle
Sludge
FIG. 32 Mow distribution baffle to increase the velocity above tubesheet, minimize the flow
stagnation zones, and minimize deposits on tubesheet (Courtesy ofP J Meyer, Siemens A G.)
150
Feedwater
Nozzle
J-Tube
Thermal Sleeve
c-Vessel Wall
Feedwater Ring
Emergency
Feedwater Nozzle
FIG. 33. Siemens/KWU feedwater distribution system with an antistmtification loop.
(Courtesy of P. J. Meyer, Siemens AG.)
the U-bends in the new steam generators have been first reduced by the use of better
manufacturing processes and then by using peening and stress relieving techniques,
respectively. The new steam generators generally contain a minimum of three sets of
antivibration bars rather than two, which are inserted deeper into the tube bundle to improve
bundle stability and reduce flow induced vibration. The bar material has been selected to
provide a better wear match to the tubing, and tube-to-antivibration bar clearances are
rrrinirnized. Also, the antivibration bars are field replaceable in some designs. The
tube-support structures in new steam generators (such as the quatrefoil, trifoil, and eggcrate
designs) prevent fluid and impurity stagnation in the tube/tubesheet annuli. Finally, new
tube/tubesheet joints in the replacement steam generators, consisting of a full length hydraulic
expansion and (for some vendors) either a one or two step hard mechanical roll, have
eliminated crevices where impurities can concentrate.
More corrosion resistant materials are being used. The US industry's consensus on the
best steam generator tube material is thermally treated Alloy 690, which is also being used
in France, Japan, and elsewhere. Alloy 800M tubing is being used in Belgium, Canada,
Germany, and Spain. The tube-support structures in new steam generators are now being
fabricated with 12% chromium ferrotic stainless steels such as Types 409, 410 or 405 to
preclude denting.
151
Other improvements include: increased blowdown capacity to help remove impurities and
reduce accumulation of sludge; a flow redistribution baffle plate to direct the recirculation
water across the tubesheet at a velocity sufficient to minimize sludge deposition on the
tubesheet (Fig. 32); a slightly reduced pitch between tubes so that the number of tubes can
be increased to allow for additional tube plugging margin or provide more power; slightly
redesigned steam generator shells to reduce the number of welds that must be inspected and
eliminate the longitudinal welds; forged heads with integrated nozzles, man-ways and support
pads; redesigned feedwater piping to prevent water hammer events and minimize stratification
in the nozzle (Fig. 33); and, a shell design modified to mitigate the girth weld cracking
problem (the girth weld is not located at the upper shell and transition cone joint but at some
distance above the joint).
Ease of future mitigation, repairs and replacements is now being considered in the design
of the replacement steam generators. Improved access for secondary-side lancing and chemical
cleaning of the tubesheet top surface has been incorporated with larger man ways and hand
holes in appropriate locations. Access for sleeving and other repairs has also been improved.
The original (older) designs of the PWR steam generators and containments did not anticipate
the need for steam generator replacement during the plant life; however, this need is
addressed in the newer designs.
8.7.2. CANDU Jteplaceirent Designs
The current design of steam generators for CANDU units has stabilized Alloy 800M
tubing, stainless steel lattice grid tube supports, flat U-bend restraints (AVBs), and high
capacity separators in an integral drum. Most CANDU steam generators are of the integral
preheater type as well.
8.7.3. WWER Replacement Designs
As mentioned in Section 2.4, the WWER-1000U steam generator has been designed to
replace the original WWER-1000 steam generators as needed The WWER-1000U has the
perforated areas of the collectors fabricated from titaniurn-stabilized austenitic stainless steel
(the same steel used for the tubes and the WWER-440 collectors), somewhat fewer tubes
(9157 versus 11 000), and a somewhat larger specific heat flux. The WWER-1000U design
details are listed in Table IV.
152
The information presented in this TECDOC suggests that steam generator tubing
degradation caused by stress corrosion cracking, fretting, vibration induced fatigue and other
age-related mechanisms continues to be a significant safety and cost concern for many steam
generators, even for some of the replacement steam generators with improved materials and
designs. Also, stress corrosion cracking, corrosion fatigue and thermal fatigue have caused and
will probably continue to cause cracking in some PWR steam generator shells and feedwater
nozzles. And, stress corrosion cracking of the WWER collector material, and possibly the
cover bolts, may also continue to occur. Therefore, systematic steam generator ageing
management programmes are needed at all nuclear power plants.
The preceding sections of this TECDOC dealt with the key elements of a steam
generator ageing management programme (AMP) whose objective should be to maintain the
fitness-for-service of the steam generators at a nuclear power plant throughout their service
life. Section 9 and Fig. 34 show how these elements are integrated within a plant specific
steam generator AMP. Such an AMP should be implemented in accordance with guidance
prepared by an interdisciplinary steam generator ageing management team organized at a
corporate or owners' group level. For guidance on the organizational aspects of a plant AMP
and interdisciplinary ageing management teams refer to IAEA Safety Practice
"Implementation and Review of Nuclear Power Plant Ageing Management Programme."
A comprehensive understanding of a steam generator, its ageing degradation, and the

effects of the degradation on the steam generators ability to perform its design functions is
the fundamental element of an AMP. This understanding is derived from a knowledge of the
design basis (including applicable codes, and regulatory requirements); the design and
fabrication (including the materials properties and specified service conditions); the operation
and maintenance history (including commissioning and surveillance); the inspection results;
and generic operating experience and research results. Sections 1.1, 2, 3 and 4 contain
information on important aspects of the understanding of steam generators and their ageing.
Section 4.4 of this TECDOC also contains a detailed summary of the steam generator
tube ruptures that have occurred to date around the world (nine tube ruptures and seven
incipient tube ruptures). The summary includes details of rupture size and location,
contributing factors, maximum leak rates and a description of the resulting consequences. In
all cases, the plants were properly cooled and radioactive releases were small and well below
regulatory limits. On the other hand, these ruptures as well as the widespread tube degradation
reported in Section 4 have been a significant concern to the safety authorities in various
countries.
In order to maintain the fitness-for-service of a steam generator it is necessary to control

within defined acceptable limits the age-related degradation of the steam generator. Ageing
degradation control consists of the following elements, based on an understanding of steam
generator ageing:
prudent operation consistent with operational guidelines aimed at rriinirnizing degradation

(Section 5);
153
Plan
2. Definition of SG Ageing
Management Programme (AMP)
Integrating ageing management activities:
• Co-ordinate AM activities
• Document AMP
Improve AMR • Optimize AMP based on current Minimize
understanding, self-assessment expected
and peer reviews degradation
Act Do
5. SG Maintenance: Mitigation, 1. Understanding SG Ageing 3. SG Operation
Repair, and Replacement
Key to effective ageing management: Managing ageing mechanisms:
Managing ageing effects:
• Materials and material properties • Follow operating guidelines
• Criteria for maintenance work • Stressors and operating conditions • Control of water chemistry,
• Mitigation/repair of tube • Ageing mechanisms impurity incursions,
degradation • Degradation sites and deposits
• Vibration control • Condition indicators • Removal of secondary
• Mitigation/repair of FW nozzles • Consequences of age-related side crevice impurities
and adjacent piping degradation and failures under
• Chemical cleaning normal operating and DBE conditions
• SG replacement
Check
4. SG Inspection, Monitoring,
and Assessment
Correct Check for
Detecting and assessing ageing effects:
unacceptable degradation
degradation • Tubing inspection
• FW nozzle, adjacent piping, shell
inspection
• Fatigue monitoring
• Leak rate monitoring
• Frtness-for-service assessment
FIG. 34. Key elements of steam generator ageing management programme (AMP) and their
interfaces. Based on Plan-Do-Check-Act elements.
inspection and monitoring consistent with requirements aimed at timely detection and
characterization of any degradation (Section 6);
assessment of the observed degradation in accordance with appropriate guidelines to
determine fitness-for-service (Section 7); and
mitigation, repair or replacement to correct unacceptable degradation (Section 8).
A steam generator AMP is a mixture of the above elements and specific ageing
management actions designed to minimize, detect and mitigate ageing degradation before the
steam generator safety margins are compromised. This mixture reflects the level of
understanding of the steam generator ageing, the available technology, the regulatory/licensing
requirements, and plant life management considerations/objective. Timely feedback of
experience is essential in order to provide for ongoing improvement in the understanding of
the steam generator ageing degradation and in the effectiveness of the AMP. The following
subsections address the main features and interfaces of key elements of a steam generator
AMP as shown in Fig. 34.
154
9.1. UNDERSTANDING STEAM GENERATOR AGEING AND FEEDBACK OF
OPERATING EXPERIENCE
Understanding steam generator ageing is the key to effective management of steam

generator ageing, i.e. it is the key to: integrating ageing management activities within a
systematic AMP, managing ageing mechanisms through prudent operating procedures and
practices; detecting and assessing ageing effects through effective and practical inspection,
monitoring and assessment methods; and managing ageing effects using proven maintenance
methods. This understanding consists of a knowledge of steam generator materials and
material properties, stressors and operating conditions, likely degradation sites and ageing
mechanisms, condition indicators/data needed for the assessment and management of steam
generator ageing, and of the consequences of age-related degradation and failures both under
normal operating conditions and design basis event (DBE) conditions.
The understanding of steam generator ageing is derived from the steam generator
baseline data, the operating and maintenance histories, and external experience. This
understanding should be updated on an ongoing basis to provide a sound basis for the
improvement of the AMP and operating, inspection, monitoring assessment and maintenance
methods and practices.
The steam generator baseline data consists of the performance requirements, the design
basis (including codes, standards, regulatory requirements), the original design, the
manufacturer's data (including materials data), and the commissioning data (including
inaugural inspection data). The steam generator operating history includes the pressure-
temperature (P-T) records, system chemistry records, and significant event reports. The steam
generator maintenance history includes the inspection records and assessment reports, design
modifications, and type and timing of maintenance performed. Retrievable, up-to-date records
of this information are needed for comparisons with applicable external experience.
External experience consists of the operating and maintenance experience of (a) steam
generators of similar design, materials of construction and fabrication; (b) steam generators
operated under similar water chemistry or with similar tube alloy material, even if the steam
generator designs are different; and (c) relevant research results. It should be noted that
effective comparisons or correlations with external experience require a detailed knowledge
of the steam generator design and operation. The present TECDOC is a source of such
information. However, this information has to be kept current using feedback mechanisms
provided, for example, by owners' groups. External experience can also be used when
considering the most appropriate inspection method, maintenance procedure and tooling.
9.2. DEFINITION OF STEAM GENERATOR AMP
Existing programmes relating to the management of SG ageing include operations,

surveillance and maintenance programmes as well as operating experience feedback, research
and development (R&D) and technical support programmes. Experience shows that ageing
management effectiveness can be improved by integrating and co-ordinating relevant
programmes and activities within a systematic ageing management programme. Safety
authorities increasingly require licensees to define such AMPs for selected systems, structures
155
and components (SSCs) by documenting relevant programmes and activities and their
respective roles in managing SSC ageing. A definition of a steam generator AMP includes
also a description of mechanisms used for programme co-ordination and continuous
improvement. The continuous AMP improvement or optimization is based on current
understanding of steam generator ageing and on results of self-assessments and peer reviews.
9.3. STEAM GENERATOR OPERATION
Operating conditions and practices significantly influence steam generator degradation,

and therefore, are the primary means for the staff of a nuclear power plant to minimize
degradation caused by potential ageing mechanisms. These practices include primary and
secondary water chemistry control, control of secondary side impurity incursions (condenser
integrity, use of condensate polishers, recycle of the blowdown water, control of lead
contamination, and removal of copper from the secondary coolant system), removal of
secondary-side crevice impurities (flushes and soaks), and control of steam generator deposits
and crud (air in-leakage, steam generator lay-up practice, pH, steam generator blowdown
system performance, and crud lancing). These activities are, of course, closely related and
were discussed in detail in Section 5. They are summarized below.
The secondary coolant water chemistry is extremely important and a secondary side
chemistry programme must be developed for the specific conditions of the plant and
maintained to minimize the corrosion of the steam generator tube and shell and the balance
of plant materials. In general, an all-volatile treatment programme is used with (a) one or
more amines to maintain cycle pH, (b) moderate to high concentrations of hydrazine to
maintain a reducing environment and scavenge small amounts of dissolved oxygen, and (c)
other additives (e.g. boric acid) to modify the crevice chemistry. The exact combination of
additives to be used will vary somewhat from plant to plant depending on such things as the
tube degradation experiences, steam generator design, makeup water chemistry, type and use
of condensate polishers, balance of plant materials and corrosion experience, hideout return
experience, etc. In addition, the water chemistry programme should limit the steam generator
secondary side water impurity concentrations to certain specified values. The programme must
include adequate sampling and specific action levels up to and including plant shutdown.
A second closely related area of plant operation is the control of secondary side impurity
incursions. This includes policies and practices associated with condenser integrity, use of the
condensate polishing system, recycling of the blowdown water, lead contamination and
removal of copper from the secondary coolant system. The condenser must be essentially
leaktight over the life of the plant. Although condensate polishers are an effective means of
achieving water purity during startup and can protect against major chemistry excursions, they
can also release low concentrations of impurities that are known corrodents. They need to be
used with care. A blowdown recovery system is used to reduce the quantity of make-up water
and the chlorides and other impurities carried by the make-up water and introduced into the
steam generator. Effective operating procedures should be used to control both acute and
chronic sources of lead contamination. And, copper-bearing alloys should be removed from
the secondary coolant system and replaced with stainless steel components.
156
A third related area of plant operation is removal of impurities from the secondary-side
crevices. As discussed in Section 4, impurities concentrate in the tubesheet and tube-to-tube
support plate crevices, the sludge pile, and under freespan crud (bridging) deposits. Steam
generator hot soaks and flushes are used, as necessary, to remove impurities from these
locations (after sludge lancing).
The fourth important and related area of plant operation is control of the steam generator
deposits (crud). Plant procedures should be in place to rninimize the leakage of air into the
secondary coolant system through the aggressive use of state-of-the-art detection and repair
technologies. Wet lay-up with chemically treated, de-oxygenated water and a positive pressure
nitrogen blanket can be used whenever possible when the steam generator is in a shutdown
condition. The feedwater pH should be between 8.8 and 9.2 for plants with copper alloys and
above 9.3 for all ferrous plants. The blowdown system should continuously remove and clean
about 1% of the main steaming rate, and up to 7% of the main steaming rate during short
transients. And, maybe most importantly, sludge lancing should be performed periodically to
remove hard crud.
9.4. INSPECTION, MONITORING AND ASSESSMENT
The steam generator inspection, monitoring and assessment activities are designed to
detect and characterize significant component degradation before the steam generator safety
margins are compromised. Together with an understanding of the steam generator ageing
degradation, the results of the steam generator inspections provide a basis for decisions
regarding the type and timing of maintenance actions and decisions regarding changes in
operating conditions and practices to manage detected ageing effects.
9.4.1. Inspection and Monitoring
Current inspection and monitoring requirements and techniques for steam generator tubes
as well as feedwater nozzles, adjacent piping and the steam generator shells are described in
Section 6. In general, the rigor and extent of the inspection increases dramatically as the
steam generator develops problems. Normally, an inspection of a significant fraction of the
tubes (and collectors or other components) with non-destructive techniques is required,
supplemented by destructive examinations such as metallography on pulled tubes.
It is extremely important to know the accuracy, sensitivity, reliability and adequacy of

the non-destructive methods used for the particular type of suspected degradation. The
performance of the inspection methods must be demonstrated in order to rely on the results,
particularly in cases where the results are used hi fitness-for-service assessments. Inspection
methods capable of detecting and sizing expected degradation are therefore selected from
those proven by relevant operating experience. Current methods used for the inspection of
steam generator tubes and their respective detection and sizing capabilities are described in
detail in Section 6.2.
157
9.4.2. Assessment
The main safety function of steam generator tubes is to act as a barrier between the
radioactive primary side and the non-radioactive secondary side. Safety margins are part of
the design and licensing requirements of a nuclear power plant to ensure the integrity of the
tubes under both normal and accident conditions. A fitness-for-service assessment is used to
assess the capability of the tubes to perform the required safety function, within the specified
margins of safety, during the entire operating interval until the next scheduled inspection.
Fitness-for-service assessments have used a variety of methods in response to the

particular conditions and circumstances present at the time of the assessment. Section 7.2 of
this TECDOC describes the fimess-for-service guidelines used in 12 different countries.
Generally, a graduated approach with three levels of assessment is used to demonstrate
fitness-for-service.
(1) The simplest and most conservative assessment method uses generic criteria (e.g. "No
flaws allowed" or "40% of wall limit") for all types of defects and degradation
mechanisms. This method is discussed in Section 7.1.1.
(2) When the generic criterion is exceeded, a degradation specific assessment can be applied.
This method reduces the conservatism of the generic criteria by using a deterministic
analysis of the specific types of degradation detected, (e.g. per ASME rules and safety
margins). It is used by plant operators to reduce steam generator repair work (e.g.
plugging) that would be required by the generic criteria. The degradation specific
assessment is described in Section 7.1.2.
(3) A probabilistic assessment may be used to demonstrate fitness-for-service when a

degradation specific assessment fails to meet the criteria. This method requires
probabilistic calculations to assess the conditional probability of tube failures, leak rates,
and ultimately risk of core damage or of exceeding site dose limits. Risk calculations
should take into account the probability that some degraded tubes may be inadvertently
left in service and that non-steam generator design basis events may increase with time
due to other age related degradation. Although probabilistic assessments of steam
generators have not been widely used, this method provides a useful way for handling
uncertainty while avoiding excessive conservatism. For more details see Section 7.4.
9.5. MAINTENANCE: MITIGATION, REPAIR AND REPLACEMENT
A variety of maintenance actions are available to manage ageing effects detected by

inspection and monitoring methods in different parts of a steam generator (see Section 8).
Decisions on the type and timing of the maintenance actions are based on an assessment of
the observed ageing effects, available decision criteria (e.g. for tube plugging), an
understanding of the applicable ageing mechanism(s), and the effectiveness of available
maintenance technologies.
Maintenance actions for managing/repair of steam generator tube degradation include

preventive/mitigation methods such as rotopeening, shot peening and stress relieving; and
158
corrective/repair methods such as plugging, sleeving and nickel plating. Tube wear/fretting
problems have been treated by preheater design modifications that reduce turbulence and peak
feedwater flow velocities and by installing additional or longer antivibration bars. Design
modifications have been used to mitigate the thermal fatigue of the feedwater nozzles and
adjacent piping.
Maintenance actions for steam generators with highly susceptible material or exposed
to very poor water chemistry may have to include certain additional measures such as molar
ratio control, chemical cleaning, and ultimately steam generator replacement.
RERERE3VCES
ASME 1992, 1992 ASME Boiler and Pressure Vessel Code, Section XI, Article IWA-4000.
Angwin, M.J. 1984. "High Temperature Pitting in PWR Steam Generator," Proceeding of the
International Congress on Metallic Corrosion, Volume 1, Toronto, Ontario, June 3—7,
1984, National Research Council of Canada, Toronto, pp. 250-256.
Azodi, D., et al. 1987. "On the Integrity of Steam Generator Tubes and Plugging
Assessment," Structural Mechanics in Reactor Technology, Transactions of the 9th
International Conference, Lausanne, August 17-21.
Bamford, W.H., G.V. Rao and J.L. Houtman 1992. "Investigation of Service-Induced
Degradation of Steam Generator Shell Materials," Proceedings of the Fifth
International Symposium on Environmental Degradation of Materials in Nuclear Power
Systems — Water Reactors, August 25-29, Monterey, California, pp. 588-595.
Bamford, W.H., et al. 1987. "Integrity Issues in PWR Steam Generator and Feedwater
Systems," Performance and Evaluation of Light Water Reactor Pressure Vessels,
PVP-Vol. 119, American Society of Mechanical Engineers, New York, pp. 19-30.
Benson, J. 1988. "Application of an F* Probe at Connecticut Yankee," presented at the EPRI

Steam Generator Workshop, Myrtle Beach, South Carolina, June 7-9.
Berge, Ph., 1993. "After 33 Years, It is Time to Solve Pure Water Cracking," Keynote speech
in Proceedings: 1992 EPRI Workshop on PWSCC of Alloy 600 in PWRs, EPRI TR-
103345, December.
Berge, Ph., et al. 1992. "The Effects of Chemical Factors on Stress Corrosion of Alloy 600
Exposed to the Cooling Medium in Pressurized Water Reactors," Proceedings of the
Fifth International Symposium on Environmental Degradation of Materials in Nuclear
Power Systems—Water Reactors, August 25-29, 1991, pp. 533-538.
Bimont, G., and G. Cordier, 1989. "A New Approach for NSSS Fatigue Life Assessment,"
Transactions of the 10th International Conference on SMIRT, Anaheim, Ca, Volume
D, pp. 37-47.
Bisbee, L.H., 1994. "Enhanced Ultrasonic Examination of Feedwater Pipe-to-Nozzle Welds,"

Nuclear Plant Journal, March-April, pp. 42-53.
Blomgren, J.C. 1986. "Experience at Zion," presented at the EPRI Workshop on Fretting and
Wear Susceptibility in Nuclear Steam Generators, Washington, D.C., March 19-20,
1986.
Bonn, L. (1988). "OTSG: Summary of Plant Experience," 7957 EPRI Workshop on

Secondary Side Intergranular Corrosion Mechanisms: Proceedings, EPRI-NP-5971,
Volume 2, pp. 4-4 to 4-5 and C4-1 to C4-11.
161
Bollini, G, 1993. "Steam Generator Inspection Reliability and Safety," Second ASME/JSME
International Conference on Nuclear Engineering (ICONE), San Francisco, March.
Brook, M.V. 1986. "Ultrasonic Transducers Design for Detection and Characterization of
Cracks in Nuclear Primary Pressure Piping," Proceedings of the 8th International
Conference on NDE in the Nuclear Industry, 17-20 November 1986, Kissimmee,
Florida, American Society for Metals, Materials Park, Ohio, pp. 439-445.
Brown, S.D. 1986. "NDE Measurement of Wear and Tube/Tube Support Plate Clearance,"
presented at the EPRI Workshop on Fretting and Wear Susceptibility in Nuclear Steam
Generators, Washington, D.C., Match 19-20, 1986.
Canadian Standards Association 1994. "Periodic Inspection of CANDU Nuclear Power Plant
Components", CAN/CSA-N285.4.
Clark, R.A., et. al. 1985. Steam Generator Group Project, Annual Report 1984,
NUREG/CR-4362, PNL 5417.
Clark, R.A., and M. Lewis 1985. "Observations of Denting and Other Deterioration in the
Surry Steam Generator," Proceedings of the International Conference in Nuclear Power
Plant Ageing, Availability Factor and Reliability Analysis, San Diego, California, July
8-12, 1985, V. S. Goel (ed), American Society for Metals, Metals Park, Ohio, pp.
193-195.
Cochet, B., 1989. Tube Plugging in the Tubesheet Area Leak Before Break Analysis - Tube
Plugging Criteria, December
Cofie N.G. et al. 1994., "Management of Steam Generator Feedwater Nozzle Cracking in
PWRs," 1994 Pressure Vessels and Piping Conference, Minneapolis, Minnesota, June
19-23, ASME PVP-Vol. 286, Changing Priorities of Codes and Standards: Failure,
Fatigue, and Creep.
Commonwealth Ediston, Co. 1987. Technical Specifications for Brddwood Station Units 1
and 2, Appendix A to License No. NPF-72, docket Nos. STN 50-456 and STN 50-
457, also published by the USNRC as NUREG-1276, July.
Cowfer, C.D., 1989. "Basis/Background for ASME Code Section XI Proposed Appendix Vfl:
Ultrasonic Examination Performance Demonstration, "Nondestructive Evaluation NDE
Planning and Application, R. D. Streit (ed.), Presented at The Pressure Vessels and
Piping Conference, Honolulu, Hawaii, July 23-27, 1989, NDE-Vol. 5, American
Society of Mechanical Engineers, New York
Cowfer, C.D., 1989. "Basis/Background for ASME Code Section XI Proposed Appendix VII:
Ultrasonic Examination Performance Demonstration," Nondestructive Evaluation NDE
Planning and Application, R. D. Streit (ed.), presented at the Pressure Vessels and
Piping Conference, Honolulu, HI, July 23-27, 1989, NDE-Vol. 5, American Society
of Mechanical Engineers, New York.
162
Dembek, S. 1995. "Summary of Electric Power Research Institute (EPRI) Lead Presentation
on Steam Generator (SG) Tube Circumferential Cracking Sizing Techniques,"
Memorandum to J. Strosnider, USNRC, March 13, 1995.
Dobbeni, D. 1991. "Ultrasonic Inspection Methodology," NEA/CSNI- UNIPEDE Specialist

Meeting on Operating Experience with Steam Generators, Brussels, Belgium, 16-20
September.
Dodd, C.V., et al. 1988. "Eddy-Current Inspection of Ferromagnetic Materials Using Pulsed
Magnetic Saturation," Material Evaluation 46, 12, pp. 1592-1597.
Dooley, B.R, et al. 1995. "Chapter 21, Ionic Impurity Control" in the EPRI Steam Generator,
Reference Book, 2nd edition, draft.
Ellison, P.O., et al. 1995. "The Risk Significance of Induced Steam Generator Tube Rupture,"
NRC/CR-XXXX draft, September.
Engstrom, J. 1985. "Comparison of Eddy Current Test Results in Steam Generator Tubesheet
Roll with Results from Metallographic Examination of Pulled Tubes," presented at the
SGOG/EPRI Workshop, St. Petersburg Beach, Florida, December 10-12, 1985.
EPRI 1995a. Steam Generator Progress Report, Rev. 11, EPRI TR-106365.
EPRI 1995b. PWR Steam Generator Tube Repair Limits - Technical Support Document for
Outside Diameter Stress Corrosion Crackingat Tube Support Plates, EPRI TR-100407,
Rev. 2A, January.
EPRI 1994. Steam Generator Progress Report, Rev. 10, Energy Management Services, Inc.,
Little Rock, Arkansas, November.
EPRI 1993d. Proceedings: 1992 EPRI Workshop on PWSCC of Alloy 600 in PWRs, EPRI
TR-103345, December.
EPRI 1993c. PWR Secondary Water Chemistry Guidelines, Rev. 3, EPRI TR-10234, Rev. 3,
May.
EPRI 1993b. PWR Steam Generator Tube Repair Limits: Technical Support Document for
Expansion Zone PWSCC in Roll Transitions, EPRI NP-6864-L, Rev. 2, August.
EPRI 1993a. PWR Steam Generator Tube Repair Limits - Technical Support Document for
Outside Diameter Stress Corrosion Cracking of Tube Support Plates, EPRI TR-
100407, Rev. 1, August.
EPRI 1990. PWR Primary Water Chemistry Guidelines: Revision 2, EPRI NP-7077,
November.
EPRI 1985. Steam Generator Reference Book, Steam Generator Owners Group, Electric
Power Research Institute, Palo Alto (proprietary report, not publicly available).
163
EPRI1982. PWR Secondary Water Chemistry Guidelines, EPRINP-2704-SR, Electric Power
Research Institute, Palo Alto.
Erdogan, F., 1976. "Ductile Fracture Theories for Pressurized Pipes and Containers,"
International Journal of Pressure Vessels and Piping, Vol. 4, pp. 253-283.
Flesch, B., and B. Cochet, 1990. "Leak-Before-Break in Steam Generator Tubes,"

International Journal of Pressure Vessels and Piping, Vol. 43, p. 165-179.
Frank, L. 1984. Steam Generator Operating Experience Update 1982-1983, NUREG-1063.
Georgia Power Co. 1987. "Steam Generator Technical Specification 3/4.4.5" in Vogtle
Electric Generating Plant Unit 1 and 2 Technical Specifications, 1987 (Unit 1) and
1989 (Unit 2), Appendix, A to Operating Licenses NPF-68 andNPF-81, Docket Nos.
50-424 and 50-425.
Giacobbe, F.S., et al. 1988. "IGA/IGSCC of Sensitized Alloy 600 Tubing in TMI Unit 1
OTSGs," 1987 EPRI Workshop on Secondary-Side Intergranular Corrosion
Mechanisms: Proceedings, EPRI NP-5971, Volume 2, Electric Power Research
Institute, Palo Alto.
Gomez Briceno, D., et al., 1994. "Stress Corrosion Cracking Susceptibility of Steam
Generator Tube Materials in AVT Chemistry Contaminated with Lead", Chemistry in
Water Reactors: Proceedings of the International Conference, Nice - France, April 24-
27, 1994.
Gonzalez, F., and P. Spekkens, 1986. "Concentration Processes under Tubesheet Sludge Piles
in Nuclear Steam Generators" in Proceedings Water Chemistry and Materials
Performance Conference, October.
Gorbatykh, V.P., 1993. "The Corrosion Life of Metal," Thermal Engineering, Vol. 40, No.
7.
German, J.A., J.E. Harris, and D.B. Lowenstein, 1995. "Steam Generator Tube Fimess-for-
Service Guidelines", AECB Report INFO-0572, My.
German, J.A., et al., 1994. Steam Generator Tube Fltness-For-Service Guidelines, Atomic
Energy Control Board, Ottawa, Canada, DEI-399, August.
German, J.A. 1989. Survey of PWR Water Chemistry, NUREG/CR-5116, ANL-88-43.
German, J.A., and E.S. Hunt 1986. Status of Cracking and Remedial Measures for PWR
Steam Generators with Full-Depth Expanded Tubing, EPRI NP-4459-LD, Electric
Power Research Institute, Palo Alto.
German, J.A., and J.A. Mundis, 1983, Workshop Proceedings: Replacement/Repair of Steam
Generators, EPRI NP-3207-SR, August.
164
Grant, I.M., 1994. "Canadian Approach to Regulation of Steam Generator Safety", Steam
Generator and Heat Exchanger Conference Proceedings, Toronto.
Harberts, C.N. 1986. "Steam Generator Tubing Wear Experience at San Onofre Units 2 and
3," presented at the EPRI Workshop on Fretting and Wear Susceptibility in Nuclear
Steam Generators, Washington, D.C., March 19-20, 1986.
Hedner, G., 1992. "Regulatory Practices and Experiences on Steam Generators," Presented
at OECD-CNRA Special Issue Meeting on Regulatory Practices and Experience on
Steam Generators, Paris, June 25-26.
Hofinann, P.J., et al. 1986. "Pressurized Water Reactor Steam Generator Tube Fretting and
Fatigue Wear Characteristics," presented at the ASME Pressure Vessels and Piping
Conference, Chicago, Illinois, Paper 86-PVP-2, American Society of Mechanical
Engineers, New York.
Hunt, E.S., and JA German 1986. Status and Suggested Course of Action for
Nondenting-Related Primary-Side IGSCC of Westinghouse-Type Steam Generators,
EPRI NP-4594-LD, Electric Power Research Institute, Palo Alto, May 1986.
IAEA 1995. Draft Report of a Consultants Meeting on Safety Issues and their Ranking for
WWER 440 Model 213 Nuclear Power Plants, WWER-SC-108, 1995-04-07.
IAEA 1993. Steam Generator Collector Integrity of WWER-1000 Reactors, Final Report of
Consultants' Meetings, Vienna, Austria, May and November, WWER-RD-057,
Distribution Restricted.
Jones, J.D., et al. 1982. TMI-1 OTSG Failure Analysis Report, GPU Nuclear Corporation,
Parsippany, New Jersey.
Kergoat, M, et al. 1994. "Fatigue Monitoring of PWR Primary Loop and Unclassified
Transients," 1994 ASME Pressure Vessels and Piping Conference, Minneapolis, MN,
PVP-Volume 278, p. 85-88.
Kobayashi, T. and D.A., Shockey 1991. Fractographic Analysis of a Crack in a Zion Steam
Generator, EPRI NP-7420, Electric Power Research Institute, Palo Alto.
Koryakin, Y.I. 1993. "SG Cracks: Counting the Costs," Nuclear Engineering International,
July, pp. 43 to 45.
Kuchirka, P.J., and J.W. Cunrangham 1986. Examination of Steam Generator Tubes
R18C53HL and R18C43HL from the Jose Cabrera (Zorita) Nuclear Power Station,
EPRI NP-4760-LD, Electric Power Research Institute, Palo Alto.
Laskowski, L.J., and M.J.B. Hudson 1986. "Recirculating Steam Generator Corrosion at
Northeast Utilities Nuclear Plants: A Review," presented at the Joint ASME/IEEE
Power Generation Conference, Portland, Oregon, October 19—23, 1986, Paper
86-JPGC-NE-4, American Society of Mechanical Engineers, New York.
165
Lemaire, P., 1943. "Current Status, 1993, of ODSCC at EdF Plants," Presented at IGA/SCC
EPRI Workshop, Minneapolis, Minnesota, October 14-15.
Lott, R.G., et al. 1992. "Primary Water Stress Corrosion Crack Growth Rates in Alloy 600
Steam Generator Tubing," Proceedings of the Fifth International Symposium on
Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors,
August 25-29, 1992, pp. 525-532.
MacDonald, P.E., et al. 1996. Steam Generator Tube Failures, NUREG/CR-6365, April.
Maeda, N. and G. Yagawa, 1991. "Some Experiences and Development in Japan Related to
Nondestructive Examination and Fracture Mechanics in Structural Integrity
Assessment of Nuclear Power Plant Components," Nuclear Engineering and Design,
131, May, pp. 329-336.
Malinowski, D.D. 1995. Presentation at the NEI/NRC Meeting re Circumferential Cracking,

February 22. Presentation slides are attached to a memo from Dembek, S. to
Strosnider, J., USNRC, March 13, 1993.
Mamet, V.A., and O.I. Martynova 1993. "Hideout in NPP Steam-Generator Boiler Water and
its Effect on the Operating Reliability of Equipment," Thermal Engineering, Vol. 40,
No. 7, pp. 497 to 502.
Martynova, O.I., and V.A, Mamet 1991. "Problems of Selecting Water Chemistry for the
Secondary Loop of a Water-Moderated Water-Cooled WWER-1000 Reactor,"
Thermal Engineering, Vol. 38, No. 7, pp. 351 to 355.
Masamori, S., et al. 1988. "R&D Status of Continuous Monitoring of Plant Components",
ANS Topical Meeting on Nuclear Power Plant Life Extension, Snowbird, UT, July 31
- August 3, pp. 344-353.
Miglin, B.P., et al., 1991. Investigation of Lead as a Cause of Stress Corrosion Cracking at
Support Plate Intersections, EPRI NP-7367-M, Babcock & Wilcox Company, June.
Myake, Y., et al. 1992. "Eddy Current Detectability for Fatigue Crack of Steam Generator
Tube," Proceedings of the llth International Conference on NDE in the Nuclear and
Pressure Vessel Industries, Albuquerque, New Mexico, April 30 - May 2, pp. 71-73.
Moles, M.D.C., et al. 1994. "Ultrasonic Measurement of Steam Generator Tube Pit Depth,"
presented at the steam generator conference, Toronto, Canada.
Monter, J.V., and G.J. Theus 1982. OTSG Tube Failures: Upper Tubesheet Corrosion Tests,
EPRI NP-2790, Electric Power Research Institute, Palo Alto.
Mostafa, S., and MB. Ramsey 1994. Unit 3 Steam Generator Feedwater Nozzle Pipe
Cracking, San Onofre Unit 3, RCE 93-024, Southern California Edison.
166
Mundis, J A 1983. "PWR Secondary Water Chemistry Guidelines-An Overview," Workshop
Proceedings: Repair Replacement of Steam Generators, Appendix C, EPRI
NP-3207-SR, Electric Power Research Institute, Palo Alto.
Nordmann, F., et al. 1983. "Experimental Investigation on Denting in PWR Steam Generators:
Causes and Corrective Actions," Journal of Engineering for Power, Transactions of the
ASME, 105, 4, pp. 755-762.
Northern States Power Co. 1985. "Steam Generator Tube Inspection, Technical Specifications
4.12-1, October 1985," in Technical Specifications for Prairie Island Nuclear
Generating Plant Units 1 and 2, Appendix A to Facility Operating License DPR-42
andDPR-60, Docket Nos. 50-282 and 50-306.
Oberg, D.L. 1983. Optical Probe for Steam Generator Tube Dent Measurement, EPRI
NP-2863, Electric Power Research Institute, Palo Alto.
Owens, C.M 1987a. "A Historical View of the Importance of the Final Anneal on Primary
Side SCC Resistance of Alloy 600 Steam Generator Tubing," Proceedings: 1983
Workshop on Primary-Side Stress Corrosion Cracking of PWR Steam Generator
Tubing, EPRI NP-5498, Electric Power Research Institute, Palo Alto, pp. 14-1 to
14-19.
Palumbo, G., King, P.J., 1992. A Thermodynamic Assessment of Environmental Precursors

for Pb-Assisted Stress Corrosion Cracking in Nuclear Steam Generators, Report No.
M92-77-P, COG No.92-328, Metallurgical Research Department, Ontario Hydro
Technologies, Inc., November 9.
Pers-Anderson, E, 1993. "PWR Vessel Head Penetration Inspections," Assuring Structural

Integrity of Steel Reactor Pressure Boundary Components, 12th SMiRT-Post
Conference Seminar No. 2, Paris, France, 23-25 August, 1993.
Pinard-Legry, G., and G. Plante, 1983. Intergranular Attack of Alloy 600: High Temperature
Electrochemical Tests, EPRI NP-3062 Electric Power Research Institute, Palo Alto.
Raussokhin, N.G., et al. 1992. "Predicting the Service Life of Thermal Power Plant with
Respect to Stress-Corrosion Cracking Conditions," Thermal Engineering, Vol. 39, No.
5, pp. 265 to 270.
Rentier, R.M., et al. 1983. PWR Model Steam Generator Corrosion Studies, EPRI NP-3138,
Electric Power Research Institute, Palo Alto.
Richards, C., 1985. "Overview of Strategic Planning Considerations," Steam Generator

Repair/Replacement Workshop, Seattle, Washington., July 25-26, Electric Power
Research Institute, Palo Alto.
Rocher, A. et al. 1994. "Investigation Into Lead Limitation in Steam Generators", Electricite
de France, Production Transport - Groupe de Laboratoires, 21, Alice Privee - 93206,
Saint-Dennis Cedex 1 - France.
167
Roussel, G., 1992. "Regulatory Practices and Experience on Steam Generators in Belgium,"
Presented at the OECD-CNRA Special Issue Meeting on Regulatory Practices and
Experience on Steam Generators, Paris, June 25-26.
Roussel, G, and P. Mgnot, 1992. "Behavior of Steam Generator Tubes in Belgium and its
Consequences," Nuclear Engineering and Design, Vol. 133, pp. 63-69.
Sakai, JC, et al. 1995. "On-Line Fatigue Monitoring System for Nuclear Power Plant,"
Proceedings of the 3rd JSME/ASME Joint International Conference on Nuclear
Engineering, April 23-27. Vol.3, ASME, pp. 1577-1582.
Sanchez-Caldera, L. 1984. Corrosion-Erosion in Steam Extraction Line of Power Stations,

Ph.D. Thesis, Massachusetts Institute of Technology.
Saudan, G., 1992. "Regulatory Practices and Experience on Steam Generator in France,"
Presented at the OECD-CNRA Special Issue Meeting on Regulatory Practices and
Shah, V.N., et al. 1992. "Assessment of Primary Water Stress Corrosion Cracking of PWR
Steam Generator tubes," Nuclear Engineering and Design, 134, pp. 199-215.
Shizuma, H., 1992. "Regulatory Practices and Experience on Steam Generators in Japan,"
Presented at OECD-CNRA Special Issues Meeting on Regulatory Practices and
Solovyev S.P., 1992. Accidents and Incidents in Nuclear Power Plants, Lecture Notes,
Obninsk.
Southern California Edison Co. 1982. "Steam Generator Tube Inspection, Technical
Specifications 3/4.4.4, February 1982," in San Onofre Unit 2 Technical Specifications,
Appendix A to License NPF-10, Docket No. 50-361, also published by the USNRC
as NUREG-0741.
Stein, A.A, and A.R Mclkee (1986). "Relationship of Annealing Temperature and
Microstructure to Primary Side Cracking of Alloy 600 Steam Generator Tubing and
the Prediction of Stress Corrosion Cracking in Primary Water," Proceedings of the
Second International Symposium on Environmental Degradation of Materials in
Nuclear Power Systems- Water Reactors, La Grange Park, Illinois, American Nuclear
Society, pp. 47.
Stoller (S.M Stoller Corporation) 1990. "Forced Shutdown—SG Inspection Ports Weld
Leakage, HAZ Crack Indication—Poor Workmanship," Nuclear Power Experience,
PWR-2, V, Steam Generators, April, p. 247.
Sudduth, A. 1986. "Experience at McGuire," presented at the EPRI Workshop on Fretting and
Wear Susceptibility in Nuclear Steam Generators, Washington, D.C., March 19-20,
1986.
168
Takamatsu, H. 1994. "Japanese SG Related R&D Program", The Kanscd Electric Power Co.,
Inc. (Japan) June.
Tatone, O.S., and R.S. Pathania 1985. "Steam Generator Tube Performance: Experience with
Water-Cooled Nuclear Power Reactors During 1982," Nuclear Safety, 26, 5, pp.
623-639.
Theus, G.J., and P.L. Daniel 1984. "Corrosion in Steam Generating Systems," Corrosion in
Power Generating Equipment, M O. Speidel and A. Atrens (eds.), Plenum Press, New
York, pp. 185-232.
Thompson, RB., and R.K. Elsley 1983. A Prototype EMAT System for Inspection of Steam
Generator Tubing, EPRI NP-2836, Electric Power Research Institute, Palo Alto.
Titov, V. F. 1993. "Causes of Damage to Cold Headers for Horizontal PGV-1000 Steam
Generators and Measures for Increasing their Operational Reliability and Lifetime,"
Thermal Engineering, Vol. 40, No. 12, pp. 959 to 966.
Titov, V.F., et al. 1992. "Damage to Heat Exchanger Tubes in Steam Generators on Nuclear
Power Stations with Water-Moderated Water-Cooled Reactors," Thermal Engineering,
Vol. 39, No. 3, pp. 158 to 160.
Titov, V.F. 1991. "Repairing and Replacing SGs at Soviet 1000 MWe PWRs," Nuclear
Engineering International, January, pp. 20 to 22.
Togo, et al. 1985. "Preventing Tube Degradation in Japan," Nuclear Engineering International,
30, 365, pp. 43-44.
Turner, AP.L. (Dominion Engineering) 1988. Private communication with V. N. Shah.
USNRC 1995a, "Circumferential Cracking of Steam Generator Tubes," NRC Generic Letter
95-03, April 28.
USNRC 1995b. "Voltage-based Repair Criteria for Westinghouse Steam Generator Tubes
Affected by Outside Diameter Stress Corrosion Cracking," NRC Generic Letter 95-05,
August 3.
USNRC 1994. "Operational Experience on Steam Generator Tube Leaks and Tube Ruptures,"
NRC mforrnation Notice 94-62, August 30.
USNRC 1993. Thermal Fatigue Cracking of Feedwater Piping to Steam Generators, NRC
Information Notice 93-20.
USNRC 1991a. "Problems with the Reliable Detection of Intergranular Attack (IGA) of
Steam Generator Tubing," NRC Information Notice 91-67, October 21, 1991.
USNRC 1991b. "Failure of Westinghouse Steam Generator Tube Mechanical Plugs," NRC
Bulletin 89-01 Supp. 2, June 28, 1991.
169
USNRC 1990. "Stress Corrosion Cracking in PWR Steam Generator Tubes," NRC
Information Notice No. 90-49, August 6, 1990.
USNRC 1990a. "Cracking of the Upper Shell-to-Transition Cone Girth Welds in Steam
Generators," USNRC Information Notice 90-04.
USNRC 1990b. "Failure of Westinghouse Steam Generator Tube Mechanical Plugs," NRC
Bulletin 89-01 Supp. 1, November 14, 1990.
USNRC 1989a. "Failure of Westinghouse Steam Generator Tube Mechanical Plugs," NRC
Bulletin 89-01, May 15, 1989.
USNRC 1989b. "Potential for Stress Corrosion Cracking in Steam Generator Tube Plugs
Supplied by Babcock & Wilcox," USNRC Information Notice 89-65.
USNRC 1982. "Cracking In the Shell-to-Transition Cone Girth Weld of a Steam Generator
at an Operating Pressurized Water Reactor," USNRC Information Notice 82-37.
USNRC 1981. Standard Specifications for Westinghouse Pressurized Water Reactors,

NUREG-0452, Revision 4, pp. 3/4 4-13 to 4-19.
USNRC 1980. "USNTC Investigation and Evaluation of Cracking Incidents in Piping in

Pressurized Water Reactors" NUREG 0691.
USNRC 1979. "Cracking in Feedwater System Piping," USNRC Bulletin 79-13, Revision 2.
USNRC 1976. "Regulatory Guide 1.121, Bases for Plugging Degraded PWR Steam Generator
Tubes," August.
USNRC 1975. "Regulatory Guide 1.83, In-Service Inspection of Pressurized Water Reactor
Steam Generator Tubes," Revision 1, July.
Van der Sluys, W.A., and W.H. Cullen 1987. "Fatigue Crack Growth of Pressure Vessel
Materials in Light-Water-Reactor Environment," Performance and Evaluation of
Light-Water Reactor Pressure Vessels, PVP 119, American Society of Mechanical
Engineers, New York, pp. 63—71.
Van der Sluys, W.A 1982. Corrosion Fatigue Characterization of Reactor Pressure Vessel
Steels, EPRINP-2775, Electric Power Research Institute, Palo Alto.
Virgil Summer FSAR "Chapter 5.0 - Reactor Coolant System," Final Safety Analysis Report,
Volume EX, South Carolina Electric & Gas Company, pp. 5.5-11 to 5.5-12.
Ware, AG., 1993, Fatigue Monitoring for Aging Assessment of Major LWR Components
(Draft), NUREG/CR-5824, EGG-2668, April 1993.
Westinghouse 1994. Technical Evaluation of Hybrid Expansion Joint (HEJ) Sleeved Tubes
Containing Indications Within the Upper Joint, WCAP-14182, August.
170
Westinghouse (Westinghouse Electric Corporation) 1991. Trojan Nuclear Plant Steam
Generator Tube Repair Criteria of Indications at Tube Support Plates, WCAP-13130,
Rev. 1, December.
Westinghouse (Westinghouse Electric Corporation) 1990. Indian Point Unit 2—Steam

Generator Inspection, Repair, and Restoration Program—Presentation to Nuclear
Regulatory Commission, June 1990, prepared for Consolidated Edison Company.
Westinghouse (Westinghouse Electric Corporation) 1989. Steam Generator Tube Plug

Integrity Summary Report, Non-Proprietary Report WCAP-12245, Revision 1,
Pittsburgh.
Wood, C., 1990. "How the Water-Chemistry Revolution is Reducing Exposures," Nuclear
Engineering International, pp. 27—30.
171
ABBREVIATIONS
AMP ageing management programme

AVB antivibration bar
ASME American Society of Mechanical Engineers
BWR boiling water reactor
CANDU Canada deuterium-uranium reactor
EDF Electricite de France
EFPY effective full power years
EPRI Electric Power Research Institute
IGA outside diameter intergranular attack
IGSCC outside diameter intergranular stress corrosion cracking
ISI in-service inspection
MRPC multifrequency rotating pancake coil
NUSS IAEA Nuclear Safety Standards
ODSSC outside diameter stress corrosion cracking
PWR pressurized water reactor
PWSSC primary water stress corrosion cracking
RSG recirculating steam generator
WWER water cooled energy reactor
172
CONTRIBUTORS TO DRAFTING AND REVIEW
Banic, M.J. Nuclear Regulatory Commission, USA

Bros, J. Tecnatom, Spain
Cizelj, L. Jozef Stefan Institute, Slovenia
Cornet, P. Electricite de France, France
Dierks, D. Argonne National Laboratory, USA
Dragunov, Y.G. OKB Gidropress, Russian Federation
Fluckiger, G. NOK-Beznau, Switzerland
Gold, R.E. Westinghouse Electric Co., USA
Grant, I.M. Atomic Energy Control Board, Canada
Ito, N. Mitsubishi Heavy Ind., Japan
Jarman, B.L. Atomic Energy Control Board, Canada
Kovacs, P. Paks NPP, Hungary
Kurakov, K.A. Ministry for Atomic Energy, Russian Federation
MacDonald, P.E. Idaho National Engineering Laboratory, USA
Maruska, C. Ontario Hydro, Canada
Matocha, K. Vitkovice, Czech Republic
Meyer, P.J. Siemens AG, Germany
Nonaka, A. Nuclear Power Engineering Corporation, Japan
Pachner, J. International Atomic Energy Agency
Remy, P.M. Electricite de France, France
Roussel, G. Nuclear Safety Institute, Belgium
Schwarz, T. Siemens/KWU, Germany
Shah, V.N. Idaho National Engineering Laboratory, USA
Sohn, G. Korean Atomic Energy Research Institute, Republic of Korea
Splichal, K. Nuclear Research Institute, Czech Republic
Technical Committee Meeting

Vienna, Austria: 5-9 September 1994
Consultants Meetings
Idaho Falls, USA: 7-10 September 1993
Vienna, Austria: 19-23 June 1995
173

Te 981 PRN PDF

Uploaded by

Copyright:

Available Formats

Te 981 PRN PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Te 981 PRN PDF

Uploaded by

Copyright:

Available Formats

IAEA-TECDOC-981

INTERNATIONAL ATOMIC ENERGY AGENCY

Orders should be accompanied by prepayment of Austrian Schillings 100,

ASSESSMENT AND MANAGEMENT OF AGEING OF MAJOR

Printed by the IAEA in Austria

2. STEAM GENERATOR DESCRIPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.1. PWR recirculating steam generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

3. STEAM GENERATOR DESIGN BASIS, FABRICATION AND MATERIALS . . . . 18

3.1. Codes and specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

4. STEAM GENERATOR DEGRADATION MECHANISMS . . . . . . . . . . . . . . . . . . . . 28

4.1. PWR and CANDU recirculating steam generator tubes . . . . . . . . . . . . . . . . . . . . 28

5. STEAM GENERATOR AGEING MANAGEMENT:

5.1. Primary coolant system water chemistry control parameters . . . . . . . . . . . . . . . . . 66

6. STEAM GENERATOR INSPECTION AND MONITORING REQUIREMENTS

6.1. Tubing inspection requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

7. STEAM GENERATOR ASSESSMENT METHODS AND FITNESS FOR

7.1. Tubing repair criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

8. STEAM GENERATOR MAINTENANCE: MITIGATION, REPAIR

8.1. Primary-side mitigation techniques for tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

CONTRIBUTORS TO DRAFTING AND REVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

The Safety Guide on In-Service Inspection (ISI) provides recommendations on methods,

Ageing management specific programmatic guidance is given in Technical Reports Series

2.1. PWR REORCULATiNG STEAM GENERATORS

/Main feedwater nozzle

Primary fluid inlet Primary fluid outlet

FIG. 2. Typicd design of a steam generator with apreheater (CSGORG 1983).

24 27 33 44' 51 A-M"* D2/D3' D4'

No of tubes 2604 3794 2604 3260 3388 4674 4578

No of row- 1 tubes 82 100 82 92 94 114 114

Tube pattern Square Square Square Square Square Square Square

Tubing heat Mill-annealed Mill-annealed Mill-annealed Mill-annealed Mill-annealed Mill-annealed Mill-annealed

Tubesheet Part-depth Part-depth Part-depth Part-depth Pan-depth Full-depth Full-depth

Tubesheet crevice 1825 18 18 18 19 or 20 18 18 75 or None None

Flow distribution None None None None None' D2 no Yes

2.2. CANDU RECIRCULATING STEAM GENERATORS

Manufacturer type B&W once- Combustion Engineering

No of tubes 4570 4864 5626 6307 15 531 8519 11 012

No of row -1 114 120 122 70 - 167 N/A

Tubing heat Thermally Mill-annealed or Thermally Thermally Mill-annealed Mill-annealed Mill-annealed

Tubesheet crevice None None None None 22 None None

Flow distribution Yes Yes Yes Broached No None Yes

Manufacturer MHI MHI MHI MHI MHI MHI MHI

Flow distribution None Yes None Yes Yes Yes Yes

"Tubesheet radial gap of 0.185 mm

FIG. 4a WWER-440 steam generator (side view). (Courtesy of Y. G. Dragunov, OKB

Parameters WWER-440 WWER-1000 WWER-

Heat exchanging tube material 08X18N10T 08X18N10T 08X18N10T

Hole array in the header perforated area

Parameters WWER-440 WWER-1000 WWER-

2.4. WWER STEAM GENERATORS

3.1. CODES AND SPECMCATIONS

3.2. FABRICATION AND MATERIALS

3.2.1. Heat Exchanger Tubes

Combustion Engineering Practice. The Combustion Engineering tubing was annealed at

3.2.2. Tube Installation in the Tubesheet

3.23. Tube Supports

HTS CTS B&Wonce-through

No. 7 tube support

No. 6 tube support