0% found this document useful (0 votes)
148 views10 pages

Overview and Lessons From More Than A Decade of Feeder Life Management

Paper from david publishing about feeder life management
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
148 views10 pages

Overview and Lessons From More Than A Decade of Feeder Life Management

Paper from david publishing about feeder life management
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 10

D

Journal of Energy and Power Engineering 7 (2013) 2164-2173

DAVID

PUBLISHING

Overview and Lessons from More Than a Decade of


Feeder Life Management
Aman Usmani1 and Zane Walker2
1. Manager and Principal Consultant Engineering Services AMEC NSS, Power and Process Americas, M5G 1E6, Ontario, Canada
2. Chemistry and Material Section Candu Energy Inc., Mississauga L5K 1B1, Ontario, Canada

Received: March 11, 2013 / Accepted: June 07, 2013 / Published: November 30, 2013.
Abstract: This paper provides an overview of the degradation in CANDUa (Canadian Deuterium Uranium) feeders in operating
plants observed from the mid 1990s onward. The degradation has been dominated by feeder wall thinning, caused by FAC (flow
accelerated corrosion) and feeder cracking. This paper summarizes the industrys response to the discovery of these two degradation
mechanisms and the methodologies, tools and technologies developed to monitor the degradation and assess the continued fitness for
service to manage the plant life. This paper identifies some of the lessons gained from more than a decade of industry effort, and
discusses how these lessons are being implemented in the refurbished and new CANDU plants. CANDUa is a registered trademark of
Atomic Energy of Canada Limited.
Key words: CANDU, feeder degradation, wall thinning, cracking, FAC.

1. Introduction
Feeder pipes form an integral part of the CANDU
(Canadian Deuterium Uranium) HTS (heat transport
system), transporting heavy water coolant to and from
the fuel channels and the inlet and outlet headers. By
virtue of their sheer number (up to 480 inlet and 480
outlet feeders) and their congested layout at the reactor
face and in the feeder cabinet, the feeders are precision
engineered, fabricated and installed components. Each
feeder is comprised of one or two tight-radius bends
and several long radius bends, and may contain a flow
orifice for controlling flow, a flow element for
measuring flow, and a swaged reducer for
accommodating changes in pipe diameter. All of these
components are welded together, resulting in several
thousands of welds throughout the feeder piping
system. The feeders are fabricated from SA-106 Grade
B carbon steel (Grade C in CANDU plants beginning
Corresponding author: Aman Usmani, Ph.D., research
fields: non-linear elasticity, seismic engineering, stress analysis
and equipment design. E-mail: [email protected].

with refurbished plants in 2008) with the exception of


flow devices which are fabricated from alloy 600
material.
Considering their complexity and the demanding
performance requirements, the feeder pipe systems
have performed, in general, very well. However,
beginning in the mid 1990s, higher than expected rates
of wall loss of outlet feeders were first reported at the
PLGS (Point Lepreau Generating Station) and
eventually confirmed at all CANDU stations [1]. Then,
in 1997, a through-wall crack was discovered in the
tight-radius bend of the S08 outlet feeder at the PLGS
[2]. Bend cracking was eventually found to be a
widespread degradation mechanism in the outlet feeder
piping system at the PLGS; bend cracking has not been
observed at any other CANDU station.
To address the discovery of wall thinning and
cracking in the feeder piping system, the Canadian
CANDU industry, including the CANDU utilities
(New Brunswick Power Nuclear, Hydro Qubec,
Bruce Power and Ontario Power Generation) and the

Overview and Lessons from More Than a Decade of Feeder Life Management

CANDU designers (Atomic Energy of Canada Limited,


and more recently Candu Energy Inc.) jointly funded
projects through the COG (CANDU Owners Group)
beginning in the 1990s. The motivation for this
initiative was to put in place strategies that would allow
each member utility to manage station operations in the
presence of these degradation mechanisms while still
meeting all nuclear safety and regulatory requirements.
Much of the early effort focused on developing a
technical understanding of the mechanisms responsible
for feeder degradation, as well as any potential
solutions that could be implemented, based on reviews
of the literature. In the case of wall thinning, wall loss
of carbon steel piping in power plants due to FAC
(flow accelerated corrosion) was known following a
loss-of-life accident involving a failed feedwater line at
the Surry Nuclear Power Plant in 1986 [3]. In the years
that followed, considerable knowledge has been
amassed to manage and, in many cases, mitigate, FAC
in the secondary side of nuclear power plants and
conventional fossil power plants [4-8]. However, the
solutions implemented in these cases were largely
based on replacing selected portions of the piping
system with more corrosion resistant materials and
operating the affected process systems with low levels
of dissolved oxygen [5, 9]. Such an approach could not
be implemented in the primary circuit of a CANDU
plant.
In the case of feeder cracking, carbon steel is
generally viewed as being resistant to this form of
degradation, although not completely immune [10]. At
the time of the first through wall cracking failure of the
S08 outlet feeder at the Point Lepreau Generating
Station in 1997, relatively little was known about the
cracking of carbon steel in relatively pure,
high-temperature water of CANDU reactor primary
circuit. It was concluded after thorough reviews of the
literature and consultations with industry experts that
the most likely mechanisms responsible for the
observed degradation were lower temperature creep
cracking and anodic stress corrosion cracking (flow

2165

accelerated corrosion and cracking of carbon steel


piping in primary water), although neither mechanism
was total consistent with the observed failures [11].
As more was learned about the mechanisms
responsible for the observed degradation, it became
apparent that although some mitigation of feeder wall
thinning and feeder cracking could be realized through
refinements in the operating water chemistry and
operational practices, elimination of the in-service
degradation was unrealistic and effort was focused on
strategies to manage the problem.
Although NDE (Non-Destructive Examination)
tools and methods existed for in-service inspection of
carbon steel piping, nearly all of these were developed
for large bore pipe. The relatively thin walls of carbon
steel feeders (< 7 mm), and their complex and highly
constrained layout, precluded the use of off-the-shelf
tools and methods (such as those inspection methods
developed by EPRI [12]). Similarly, Fitness for Service
Guidelines were developed specifically to deal with the
relatively thin-wall CANDU feeder pipes for
dispositioning all forms of degradation found, and in
the cases where the feeders failed to meet the
acceptance criteria, tools for the repairing and
replacing feeder components were developed. This
paper provides a summary of these activities and
demonstrates how the industry has successfully
managed feeder degradation in the current CANDU
fleet.

2. Background to Feeder Wall Thinning and


Feeder Cracking
By the mid-1990s, it became apparent that relatively
large deposits of magnetite were being observed in the
cold leg of the steam generators, and carbon steel
feeder pipes were suspected as being the source of the
iron. Wall thickness inspections performed on the tight
radius bends near the reactor face, first at PLGS in
1995 and subsequently at other CANDU stations,
revealed higher than expected rates of wall loss. The
mechanism responsible for the wall thinning of outlet

2166

Overview and Lessons from More Than a Decade of Feeder Life Management

feeders was confirmed to be flow accelerated corrosion


following the removal of a section of the S08 outlet
feeder from PLGS in 1997 and the appearance of the
characteristic scalloped surface and thin oxide
(magnetite) film on the inside surface.
With the collection a large number of wall thickness
measurements from CANDU outlet feeders over the
ensuing years, stations confirmed that [11, 13]:
Outlet feeders were experiencing a linear rate of
feeder wall loss with time;
The rate of wall loss was dependent upon coolant
velocity and turbulence, as manifested through bend
geometries and hydraulic conditions.
Not only did the development of this understanding
reinforce the conclusion that the higher than expected
wall loss was attributed to FAC, it also provided the
means to confidently predict the future wall thickness
of outlet feeders. The understanding of the roll of water
chemistry on FAC provided an opportunity to reduce
the wall thinning rate through tighter control of coolant
alkalinity [14]. Inspections performed on a significant
number of inlet feeders across the industry
demonstrated the absence of this form of degradation.
In recent years, the removal of inlet feeders during
refurbishment related activities confirmed that iron
oxides were indeed showing net deposition, rather than
net removal, at inlet feeder surfaces.
The first indication of cracking in CANDU feeders
came following the discovery of a through-wall crack
in the S08 outlet feeder bend at PLGS in 1997.
Following the discovery of an additional outlet feeder
bend with a through-wall crack, and two outlet feeder
bends with partial through-wall cracks, in 2001, the
widespread nature of bend cracking at PLGS became
apparent [11]. In the years that followed, partial
through wall cracks were detected, through in-service
inspections, in an additional eight outlet feeders. By the
time that PLGS was shut down for refurbishment, a
total of 12 outlet feeder bends had been removed due to
the presence of a mixture of inside and outside surface
initiated cracks.

In 2003 at the Gentilly-2 Nuclear Generating Station,


a crack was found in the G09 feeder field weld, which
had been repaired during construction. Post-removal
examination of the cracked weld revealed similarities
with the bend cracks observed at the PLGS, namely
that all cracks initiated in locations of Ref. [11]:
high-residual tensile stress [15] (the PLGS feeder
bends had not been stress relieved during fabrication);
in material that had been subjected to cold work.
The fact that the cracking had exclusively occurred
in outlet feeders indicated that the higher temperatures
synonymous with outlet conditions were also likely a
prerequisite for cracking susceptibility. Evidence
obtained through post-removal examination of cracked
feeder bends, as well as through laboratory testing of
feeder pipe material, indicated that the bend cracking
was most likely caused by either or both stress
corrosion cracking, in the case of the inside surface
initiated cracks and low temperature creep cracking,
possibly enhanced by the presence of atomic hydrogen
in the steel, which is formed during the FAC of the
inside surface.
In spite of extensive inspections of non-stress
relieved tight-radius bends and of feeder welds in other
CANDU plants, no evidence of additional cracking has
been observed. Although this suggests that the PLGS
feeder pipes were predisposed to cracking, no material
or operating condition unique to PLGS has yet been
identified.

3. Industry Response
Upon the discovery of feeder wall thinning in the
1990s, the Feeder Wall Thinning Project was initiated
and funded through the COG. Much of this work was
focused on improving the understanding of the extent
of feeder wall thinning in the stations and developing
the first Fitness for Service Guidelines to deal with the
observed wall loss degradation. A similar joint project
was initiated following the discovery of the first crack
in the S08 outlet feeder bend at PLGS, with an
emphasis on identifying the mechanism(s) responsible.

Overview and Lessons from More Than a Decade of Feeder Life Management

2167

With the discovery of feeder wall thinning and


subsequent feeder cracking, the CANDU industry
performed a significant amount of stress analyses using
a range of methodologies including linear,
elastic-plastic, limit and collapse analyses.
Probabilistic assessments of the degradation and extent
of condition were also performed. Effort was spent to
validate these methods and demonstrate the availability
of ample margins. The highlights of these
methodologies are provided below.

of the design and construction code. Various modeling


approaches were employed and included:
Single feeder models, as well as more complex
multi-feeder models, using uniformly thinned pipe
bends to simulate thinned pipe bend profiles with
bounding amounts of ovality;
The stress indices used in the analysis included
both code prescribed values for pipe bends and elbows
as well as more realistic, finite element calculated
values;
Elastic-plastic fracture mechanics methods;
Seismic
analysis using both linear modal
superposition floor response spectrum and direct
integration time history approach.
The industry directed effort at standardizing the
analysis methodologies, eventually producing a
document describing the various methodologies used
in each submission to the CNSC by the member
utilities. Along with a check list, this stress analysis
summary report facilitated regulatory reviews of the
fitness for service submissions. In some cases elastic
plastic, limit and collapse load methodologies of
ASME Code NB 3200 were also used to demonstrate
the fitness for service of thinned feeders. A
representation and a successful submission to revise
the number of cycles and the permissible value of
damping in seismic analysis of feeders was made to the
CSA N289 technical committee, with the revised
values appearing in the latest revision of the CSA
N289.3-10 Standard.
In parallel with the conventional ASME code
approach to demonstrate the fitness for service of
degraded feeders, the Fitness for Service Guidelines for
feeders subjected to various degradation mechanisms
were also developed. The development of the technical
basis and the FFSG is described in a subsequent
subsection.

4.1 Stress Analysis

4.2 Technical Basis Document and FFSG

The stress analysis of degraded feeders was initially


based on the ASME Section III NB3600 requirements

Over the period of a decade, more than two hundred


technical reports on various relevant topics were

Following the discovery of three cracked feeder bends


at PLGS in 2001, and the realization that the extent of
wall loss in some CANDU plants was likely to limit the
operating life of some of the outlet feeders, the FIJP
(Feeder Integrity Joint Project) was initiated with a
mandate to address all forms of feeder degradation.
The FIJP was comprised of multi-disciplinary
technical specialists organized and directed by the
Feeder Integrity Steering Committee. The technical
specialists were aligned under teams addressing:
chemistry related aspects of feeder degradation;
non-destructive examination technology;
integration of industry data and understanding;
repair and replacement technology;
fitness for service.
Strategies were developed to manage feeder
degradation, together with the tools necessary to
implement these strategies. These tools were
comprised of:
assessment
methodologies, including FFSG
(Fitness for Service Guidelines);
NDE inspection tools for performing in-service
inspections;
tools required for the repair and replacement of
feeder components.

4. Assessment Methodologies

2168

Overview and Lessons from More Than a Decade of Feeder Life Management

prepared by FIJP investigators. The results from these


technical reports ware integrated in two Technical
Basis Documents and provided the foundation for the
FFSG.
The development of FFSG was a major
accomplishment of the FIJP. The FFSG document for
CANDU feeders is similar in approach to ASME
Section VI for the operating light water reactors. The
FFSG address all three forms of degradation; namely
wall thinning, cracking/planar flaws and blunt flaws.
The guidelines provide reasonable assurance, using
industry acceptable consistent methodologies, that
feeder structural integrity is maintained and that any
consequential leakage from feeder flaws would be
acceptable in that the leak could be detected in a timely
manner and the reactor safely shutdown, if required.
The FFSG Revision 2 was accepted by the CNSC on
a trial basis for two years in 2010. The updated
Revision 3 of the FFSG will be submitted to CNSC for
approval this year.
4.3 Feeder Bend Testing
One of the major accomplishments of the FIJP was
to perform feeder bend tests simulating thinning and
cracking, under internal pressure and moments loading.
At the outset, the margins to failure of feeders
subjected to active degradation were not known. It was
recognized early on that full size components tests
would be required to take full advantage of these
unidentified margins. These tests were performed
using removed ex-service, archived and new fabricated
feeder bends. The feeder bend test program results
demonstrated that even the most significantly degraded
feeders have adequate margins to failure.
4.4 Probabilistic Methods and Software Tools
Due to some of the limitations associated with
inspecting the complex feeder piping system in a
radiation environment, the amount of inspection data
available often leads to overly conservative outcomes
when using deterministic analysis methods. As a

consequence, probabilistic methods were developed to


provide the required degree of confidence when
estimating wall thinning rates, optimizing inspection
sample size and in assessing the possible extent of
degradation in the non-inspected feeder population.
The first statistical/probabilistic approaches were
applied to feeder thinning at Pickering A NGS and to
feeder cracking at PLGS [16]. Eventually, these
probabilistic approaches were incorporated into the
following qualified software tools:
FIRST (Feeder Inspection Regression Statistical
Tool) for feeder thinning;
Pro-FAST (Probabilistic Feeder Analysis Tool)
for feeder cracking.
Probabilistic methods have also been applied to
estimate the rupture frequency of elbows with blunt
flaws [17].
4.5 Dissimilar Metal Welds
Following the OPEX (Operating Experience) from
pressurized water reactors indicating the susceptibility
of dissimilar metal welds involving nickel-based
materials, work was, under the COG R & D performed
program, to investigate the potential susceptibility of
dissimilar metal welds between the carbon steel feeder
pipes and alloy 600 flow devices located in some of the
outlet feeders at the Darlington and Bruce Nuclear
Generating Stations. CSA N285.4 standard governing
the in-service inspections of the feeder piping system
requires that locations susceptible to in-service
cracking be inspected. Unfortunately, these feeder
dissimilar metal welds are very difficult to access and
any attempts to inspect them will be dose intensive.
To manage the issue of dissimilar metal welds and
their in-service inspection requirement, probabilistic
and deterministic leak-before-break assessments were
initiated under the FIJP. These assessments are
intended to be used to demonstrate an acceptably low
probability of rupture and to demonstrate that the
margins on crack size, load and detectable leak rate are
sufficient to conclude that leakage from through-wall

Overview and Lessons from More Than a Decade of Feeder Life Management

circumferential cracks would be detected and the unit


shutdown prior to the feeder rupturing. This work was
recently transferred to the COG R & D program where
activities required to resolving outstanding issues are
underway.

5. Inspection Tools
Monitoring the degradation being experienced in the
outlet feeders of CANDU plants was complicated by
several factors unique to CANDU feeders. The
complex feeder geometry and the close proximity of
feeders to other feeders, components and support
structures, resulted in significantly challenging
clearance issues for traditional NDE tooling. The
relatively thin wall of feeder pipes meant that wall
thickness inspection tooling had to detect relatively
small changes in wall thickness with high spatial
resolution. In terms of volumetric inspection for feeder
cracks, the tooling had to have the sensitivity to detect
relatively small (shallow) cracks while maintaining a
low rate of false positive calls. The relatively high
radiation fields associated with the reactor face meant
that the inspection tooling had to operate efficiently so
as to limit the inspectors exposure to dose. These
restrictions necessitated the development of feeder
specific inspection tooling and procedures.
Several tools were developed to facilitate the wall
thickness inspection of the following feeder locations:
Feeder bendsA scanning tool [18], comprised
of 14 conventional ultrasonic probes mounted in a
flexible bracelet covering a significant fraction of the
pipes circumference, was developed. The tool can be
translated over long lengths of feeder pipe, either
manually or using a mechanized crawler, including
over bends. Data are collected simultaneously from all
14 probes with high axial spatial resolution.
Feeder pipe adjacent to weldsTo provide wall
thickness inspection data in regions immediately
adjacent to welds, such as the hub-to-pip weld, which
the 14 probe tool can not access, specialized tools
containing 6 to 8 conventional ultrasonic probes, were

2169

developed.
Feeder
pipe under welds: To provide
measurements of wall thickness directly under weld
caps, a tool based on phased-array ultrasonic
techniques, was developed.
Furthermore, to efficiently process the very large
quantities of ultrasonic time of flight data generated
with these specialized inspection techniques, the
FAATS (Feeder Automated Analysis and Trending
Software), data analysis software package was also
developed.
Specialized tools and procedures were also
developed for detecting cracks through volumetric
inspections of the following feeder locations:
Feeder bends: A manual shear wave technique
and a highly evolved procedure optimized for axial
crack detection has been widely used in the industry for
over a decade [19]. In recent years, the delivery and
application of this shear wave technique has been
automated, significantly reducing the time required for
an inspector to spend at the reactor face, thereby
reducing the associated dose.
Feeder
welds: An phased-array ultrasonic
technique and procedure were developed to detect for
the possible presence of circumferential cracks. The
delivery and application of this inspection
methodology was also automated to reduce inspector
dose.
To ensure the reliability of these NDE inspection
methodologies, and to comply with regulatory
requirements, these inspection technologies are being
qualified to the requirements of the CANDU
Inspection Qualification Bureau.

6. Feeder Repair and Replacement


Upon recognizing the significance and potential
impact of feeder wall thinning and feeder cracking, and
the likely need to replace a significant number of
feeders, effort was directed at developing tooling that
could be used to either repair feeders in-situ or to assist
with the removal of the feeder and the replacement

2170

Overview and Lessons from More Than a Decade of Feeder Life Management

with a new spool piece. Due to the limited access


associated with the feeders, especially at the reactor
face, the tools, as well as the procedures, had to be
highly customized.
To date, all feeder sections failing to meet
acceptance criteria have been cut out and replaced
using specialized tooling using procedures developed
by the service providers performing the replacements.
The FIJP also funded work to develop technology that
could provide isolation to an outlet feeder nozzle, a
location that would be impossible to isolate using
standard freeze plugging technology.

7. Lessons Learned
Upon discovering the presence of active feeder
degradation mechanisms, the industry undertook a
concerted effort to understand the mechanisms and
causes of degradation and to implement measures to
monitor and manage degradation, including tools and
methodology development. Through more than a
decade of industry experience, lessons have been
gained and are discussed below.
7.1 OPEX
Ongoing attention to operating experience from
CANDU and LWR plants and ongoing findings from R
& D is extremely important. Some examples of
relevance to feeder degradation included:
The catastrophic failure of a high pressure
condensate line in Surry Nuclear Power Plant in 1986
gave the first indication of the FAC in carbon steel pipe;
this was nearly a decade prior to the first detection of
FAC in CANDU feeders. The effectiveness of small
additions of chromium to carbon steel was fully
established and was implemented immediately in new
CANDU reactors subsequent to the discovery of feeder
wall thinning. As well a recommendation to operate the
reactors coolant circuit at the lower end of the
permissible range of pHa was also implemented.
The role of residual stresses in crack initiation and
propagation has long been known. Although stress

relieving heat treatments of cold worked feeder


material had been adopted at many CANDU plants
prior to the discovery of the first crack at PLGS, it is
now a mandatory requirement for refurbished and new
plants.
Dissimilar metal welds, involving nickel-based
materials, have been of concern to the pressurized
water reactors for many years and this OPEX was the
trigger for assessing the condition of dissimilar metal
welds in some outlet feeders at a small number of
CANDU plants.
7.2 Joint Projects
Given the relatively smaller size of the CANDU
industry, the role and effectiveness of pooling of
resources to cooperatively handle issues such as those
in feeders in operating plants is highly recommended.
The success of the COG FIJP to effectively manage
the feeder life in the operating CANDU plants is a good
example of this recommendation.
7.3 Communication with Stake Holders including the
Regulator
Information sharing, on-going communication
amongst the various stake holders and the regulatory
organization(s) is extremely valuable in avoiding
surprises and delays.
The FIJP developed FFSG, together with the
technical basis documents, and acceptance by the
CNSC is a good example of the cooperation amongst
all the stake holders. Likewise, incorporation of the
seismic number of cycles and damping into the CSA
N289.3 Standard, to reflect industry practice and
current knowledge in the assessment of the degraded
feeders, is another good example of such collaboration.
7.4 Testing
Notwithstanding advances and sophistications in
computerized analytical methods, experimental testing
is very valuable in validating the analytical techniques
and in providing reliable determination of margins to

Overview and Lessons from More Than a Decade of Feeder Life Management

failure.
7.5 Probabilistic Methods
The probabilistic methods are valuable in accounting
for uncertainties, extending the results to un-inspected
population, optimizing the inspection scope and
providing associated confidence levels as one of the
inputs to the decision makers.
7.6 Independent Reviews
To provide a high degree of confidence, it is
advisable to engage knowledgeable and reputable
independent reviewers to provide their input and
endorsement.

8. Improvements in the Refurbished and


New Designs
Considerable insights into the mechanisms of feeder
wall thinning and feeder cracking were gained through
the work performed in the FIJP, as well as through
initiatives undertaken by AECL and the CANDU
utilities. This work included the identification of key
factors that influence the extent to which these
degradation mechanisms are active in outlet feeders of
CANDU plants. Understanding how these factors
affect feeder degradation provides an opportunity to
implement new requirements during the refurbishment
and design of current and new reactors respectively to
ensure that the design life of feeders is not limited by
these mechanisms in the future.
The key improvements introduced to mitigate the
impact of feeder wall thinning include:
A
minimum chromium content of 0.30
wt%Since the FAC of carbon steel is characterized
by the rapid dissolution of the magnetite corrosion film,
the addition of even small amounts of chromium can
significantly reduce the solubility of this corrosion film,
and thereby reduce the feeder wall thinning rate [20].
Increased wall thickness of 2-inch feedersThe
corrosion allowance of 2-inch outlet feeders was
typically ~2/3 of that of 2-inch outlet feeders. As

2171

result, 2-inch outlet feeders were typically life-limiting


in CANDU 6 plants. By increasing the wall thickness
of the 2-inch pipe, the operating life of these feeders is
significantly increased.
Tighter controls on the reduction of wall
thicknessDuring feeder fabrication, the wall
thickness of a feeder can be reduced during the bending
process or as a result of localized grinding to facilitate
fit up to other components such as hubs. Requirements
have been introduced to more tightly control these
activities, to ensure that any reduction in wall thickness
is within acceptable limits and is properly documented.
Material specified to SA-106 Grade C rather than
Grade BAlthough the tensile properties of many
heats of the feeder pipe ordered for original CANDU
plants met the higher strength requirements of SA-106
Grade C material, the higher strength could not be
credited since the material was specified to Grade B
requirements. Feeder pipe is now specified to the
Grade C requirements, allowing the higher strength
properties to be credited and providing additional
margin in the feeder design.
Additional improvements have been introduced to
eliminate the susceptibility to feeder cracking in the
future.
Stress relieving of all bends and swagesSince
the cracking observed to date has all occurred in
locations of high residual tensile stress, applying stress
relieving operations to all bends will essentially
remove the driving force for crack initiation and
propagation. The stress relieving operations also
reduce the impact of changes in the microstructure of
the steel introduced during the cold forming and
welding operations, resulting in a material less
susceptible to cracking.
Killing the steel with aluminumOne of the
mechanisms believed to be responsible for feeder bend
cracking is low temperature creep cracking. Since this
mechanism is prevalent in material with high free
nitrogen content, killing the steel with aluminum
during the steel making process essentially eliminates

Overview and Lessons from More Than a Decade of Feeder Life Management

2172

free nitrogen from the steel and the susceptibility to this


form of in-service degradation.
Elimination of localized weld repairsSince
localized repaired welds, in which material is locally
removed and back filled with weld metal, are known to
be more susceptible to in-service degradation, the
elimination of such repairs in new feeders significantly
reduces the likelihood of weld cracking.

[8]

[9]

[10]

9. Summary and Conclusions


The life management of CANDU feeders over more
than a decade has provided a good understanding of the
degradation mechanisms, means to monitor, assess and
establish the fitness for service, as well as the
improvements required to mitigate these forms of
degradation in the refurbished and new feeder designs.
The lessons gained also highlight the significance of
ongoing attention to OPEX, the value of industry
cooperation, the effectiveness of information sharing
and involvement of all stake holders in tackling generic
industry issues.

References
[1]

[2]

[3]

[4]

[5]

[6]

[7]

K.A. Burrill, E.L. Cheluget, Corrosion of CANDU outlet


feeder pipes, in: Proceedings of the JAIF International
Conference on Water Chemistry in Nuclear Power Plants,
Kahiwaszaki, Japan, 1998.
A. Celovsky, M.D. Wright, T.S. Gendron, S.A. Usmani, J.
Slade, S08: Investigation and repair of a cracked feeder at
Point Lepreau, in: 4th Int. Conf. CANDU Maintenance,
Toronto, Canada, 1997, pp. 97-109.
P.C. Wu, Erosion/Corrosion-Induced Pipe Wall Thinning
in U.S. Nuclear Power Plants, US NRC Report
NUREG-1344, 1989.
B. Chexal, J. Horowitz, R. Jones, B. Dooley, C. Wood, M.
Bouchacourt, et al., Flow-Accelerated Corrosion in Power
Plants, EPRI Report TR-106611, Pleasant Hill, Calif.,
USA, 1996.
B. Dooley, Flow-accelerated corrosion in fossil and
combined cycle/HRSG plants, PowerPlant Chemistry 10
(2) (2008) 68-89.
J. Horowitz, Investigation into Flow-Accelerated
Corrosion at Low Temperatures, EPRI Report 1013474,
2006.
J. Horowitz, Chemistry Effects on Flow-Accelerated
Corrosion Pressurized Water Reactors: Hydrazine and

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

Oxygen Investigations, EPRI report, 2005.


D.H. Lister, L. Liu, A. Feicht, M. Khatibi, W. Cook, K.
Fujiwara, et al., A fundamental study of flow-accelerated
corrosion in feedwater systems, PowerPlant Chemistry 10
(11) (2008) 659-667.
W. Ruehle, H. Neder, R. Holz, V. Schneider, Oxygen
injection into reheating steam moisture separator reheaters,
in: Proceedings of the 14th International Conference on
the Properties of Water and Steam, Kyoto, Japan, Aug.
29-Sept. 3, 2004, pp. 573-580.
H.P. Seifert, S. Ritter, Research and Service Experience
with Environmentally-Assisted Cracking in Carbon and
Low-Alloy Steels in High Temperature Water, SKI report
2005:60, 2005.
J.P. Slade, T.S. Gendron, Flow accelerated corrosion and
cracking of carbon steel piping in primary water operating
experience at the Point Lepreau Generating Station, in:
Proceedings of the 12th International Conference on
Environmental Degradation of Materials in Nuclear Power
SystemWater Reactors, Salt Lake City, 2005, pp.
773-782.
R.B. Dooley, R. Tilley, Guidelines for Controlling
Flow-Accelerated Corrosion in Fossil and Combined
Cycle Plants, EPRI report 1008082, 2005.
J. Pietralik, B.A.W. Smith, CFD application to
flow-accelerated corrosion in feeder bends, in:
Proceedings of 14th International Conference on Nuclear
Engineering, Miami, Florida, USA, July 17-20, 2006.
A.J. Elliot, M.S. Godin, Z.H. Walker, In-Reactor loop
experiments to study the impact of coolant pH on the
flow-accelerated corrosion of carbon steels with varying
chromium concentrations under CANDU HTS conditions,
in: Proceedings of the International Conference on Water
Chemistry of Nuclear Reactor Systems, San Francisco,
2004, pp. 1446-1454.
M. Yetisir, S. Khajehpour, Y. Ding, Residual stresses in
CANDU feeder bendscomparison of measurements and
predictions, in: Proceedings of the 2007 ASME Pressure
Vessels and Piping Division Conference, San Antonio,
Texas, 2007.
X. Duan, M.J. Kozluk, T, Gendron, J. Slade, Crack
alternative methodology for assessing part-through-wall
cracks in carbon steel bends removed from Point Lepreau
Generating Station, Nuclear Engineering and Design 241
(3) (2011) 630-637.
M. Wang, M. Li, E. Lehockey, X. Duan, M.J. Kozluk,
Determination of rupture frequency for feeders subject to
flow accelerated corrosion, Nuclear Engineering and
Design 241 (6) (2011) 2234-2239.
E. Lavoie, G. Rousseau, L. Reynaud, On the development
of the METAR family of inspection tools, in: Proceedings
of the 6th CNS International Conference on CANDU

Overview and Lessons from More Than a Decade of Feeder Life Management
Maintenance, Toronto, Ontario, Canada, Nov. 16-18,
2003.
[19] E. Lavoie, G. Rousseau, E. Lavelle, J.F. Cot, Crack
detection tools for feeder inspection, in: Proceedings of the
7th CNS International Conference on CANDU
Maintenance, Toronto, Ontario, Canada, Nov. 20-22, 2005.

2173

[20] Z.H. Walker, B. Rankin, Benefit of chromium in reducing


the rates of flow accelerated corrosion of carbon steel
outlet feeders in CANDU Reactors, in: Proceedings of the
13th International Conference on Environmental
Degradation in Nuclear Power Systems, Whistler, Canada,
Aug. 2007.

You might also like