ML13263A271
| ML13263A271 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 09/20/2013 |
| From: | Reynolds S NRC Region 4 |
| To: | Peter Dietrich Southern California Edison Co |
| Lantz R | |
| References | |
| CAL 4-12-001, EA-13-083 IR-12-009 | |
| Download: ML13263A271 (92) | |
See also: IR 05000361/2012009
Text
September 20, 2013
CAL 4-12-001
Mr. Peter Dietrich
Senior Vice President and
Chief Nuclear Officer
Southern California Edison Company
San Onofre Nuclear Generating Station
P.O. Box 128
San Clemente, CA 92674-0128
SUBJECT: SAN ONOFRE NUCLEAR GENERATING STATION - NRC CONFIRMATORY
ACTION LETTER RESPONSE INSPECTION 05000361/2012009 AND
Dear Mr. Dietrich:
Following the June 7, 2013, announcement of Southern California Edisons decision to
permanently shut down San Onofre Nuclear Generating Station, Units 2 and 3, the U.S. Nuclear
Regulatory Commission (NRC) terminated our review of your Confirmatory Action Letter
Response (ML12285A263) for Unit 2, dated October 3, 2012. The enclosed report documents
the NRC assessment of your activities through June 7, 2013, in response to our March 27,
2012, Confirmatory Action Letter (ML12087A323). The NRC also reviewed the two remaining
open unresolved items identified in Augmented Inspection Team Report 05000361/2012007 and
05000362/2012007 (ML12188A748). The two unresolved items were related to the mechanistic
cause of the excessive and unexpected wear in both Units 2 and 3 steam generator tubes,
which resulted in a steam generator tube leak on Unit 3 on January 31, 2012. The results of
this inspection were discussed with you and other members of your staff on August 28, 2013.
The inspectors examined activities conducted under your license as they relate to safety and
compliance with the Commissions rules and regulations and with the conditions of your license.
The inspectors reviewed selected procedures, documents, and records and interviewed
personnel.
On June 12, 2013, Southern California Edison submitted a Certification of Permanent Cessation
of Power Operations letter to the NRC, certifying that Units 2 and 3 have permanently ceased
power operations. On June 28 and July 22, 2013, Southern California Edison certified that all
fuel had been permanently removed from the Units 3 and 2 reactors, respectively
(ML13183A391 and ML13204A304).
UNITED STATES
NUCLEAR REGULATORY COMMISSION
RE G IO N I V
1600 EAST LAMAR BLVD
ARLINGTON, TEXAS 76011-4511
P. Dietrich
- 2 -
When Southern California Edison announced the permanent shutdown of both Units 2 and 3,
the NRC halted its review of the operational assessments and other open issues related to the
Unit 2 CAL response, and no final determination on the adequacy of that response was made.
The NRC Office of Nuclear Reactor Regulation was also conducting a technical evaluation,
which could not be completed because further information, including the adequacy of using
squeeze film dampening for determining the effectiveness of the anti-vibration bars and
subsequent vibration response, was required. Questions from NRC inspectors about the use of
squeeze film dampening for the anti-vibration bar configuration at low frequency resulted in
additional testing being conducted at the Atomic Energy Canada Limited facility in Chalk River,
Canada, and at Mitsubishi Heavy Industries (Mitsubishi) in Kobe, Japan. Southern California
Edison determined that revisions to their previously submitted operational assessments for
Unit 2 were needed based on the testing results; however, those revisions were not completed
before the permanent shutdown announcement was made.
The enclosed report documents two NRC-identified findings associated with the thermal-
hydraulic unresolved item, one finding of very low safety significance (Green) for Unit 2 and one
finding that was preliminarily determined to have a low to moderate safety significance (White)
for Unit 3. The Mitsubishi FIT-III thermal-hydraulic computer model (FIT-III) output gap
velocities were not appropriately modified for triangular pitch designed steam generators. There
were opportunities to identify this error during the design of the replacement steam generators.
Mitsubishi was the vendor selected by Southern California Edison to design and manufacture
the replacement steam generators. On numerous occasions during the design process,
Southern California Edison personnel questioned the results from and appropriateness of using
FIT-III, but ultimately accepted the design as proposed by Mitsubishi. Mitsubishi hired
consultants with expertise in designing large steam generators, but did not rigorously evaluate
all concerns raised by the consultants about use of FIT-III and specific results obtained from that
thermal-hydraulic model. As a result, replacement steam generators were installed at San
Onofre with a significant design deficiency, resulting in rapid tube wear of a type never before
seen in recirculating steam generators. The NRC assessed these findings based on the best
available information using the applicable Significance Determination Process. For Unit 2, all
the steam generator tubes were determined to meet the technical specification requirements for
tube integrity; therefore, the design control violation for Unit 2 was determined to be of very low
safety significance. For Unit 3, we conducted an independent risk analysis and determined that
the risk was of low to moderate safety significance (White). The NRCs preliminary significance
was based on the following conservative assumptions: an exposure time of 172 days; a steam
generator tube rupture that results in core damage will always result in a large early release;
degraded tubes resulted in an increased frequency of a steam generator tube rupture; and a
main steam line break could have occurred during the exposure period, resulting in one or more
tubes rupturing. The details of all primary assumptions associated with the preliminary
significance determination are documented in Attachment 5 of the enclosed report.
As a corrective action, your staff revised the thermal-hydraulic code of record and ensured that
the code was in accordance with ASME guidance
The Unit 2 finding was determined to involve a violation of NRC requirements. The NRC is
treating this violation as a noncited violation (NCV) consistent with Section 2.3.2.a of the
P. Dietrich
- 3 -
Enforcement Policy. The Unit 3 finding is an apparent violation of NRC requirements and is
being considered for escalated enforcement action in accordance with the NRC Enforcement
Policy. The Enforcement Policy is included on the NRCs Web site at http://www.nrc.gov/about-
nrc/regulatory/enforcement/enforce-pol.html.
In accordance with NRC Inspection Manual Chapter 0609, Significance Determination Process,
we intend to complete our evaluation using the best available information and issue our final
determination of safety significance within 90 days of the date of this letter. The significance
determination process encourages an open dialogue between the NRC staff and the licensee;
however, the dialogue should not impact the timeliness of the staffs final determination.
Before we make a final decision on this matter, we are providing you with an opportunity: (1) to
attend a Regulatory Conference where you can present to the NRC your perspective on the
facts and assumptions the NRC used to arrive at the finding and assess its significance, or
(2) submit your position on the finding to the NRC in writing. If you request a Regulatory
Conference, it should be held within 30 days of the receipt of this letter and we encourage you
to submit supporting documentation at least one week prior to the conference in an effort to
make the conference more efficient and effective. The focus of the Regulatory Conference is to
discuss the significance of the finding, not necessarily the root cause(s) or corrective action(s)
associated with the finding. If a Regulatory Conference is held, it will be open for public
observation. If you decide to submit only a written response, such submittal should be sent to
the NRC within 30 days of your receipt of this letter. If you decline to request a Regulatory
Conference or submit a written response, you relinquish your right to appeal the final
significance determination, in that, by not doing either, you fail to meet the appeal requirements
stated in the Prerequisite and Limitation sections of Attachment 2 of Inspection Manual
Chapter 0609.
Please contact Ryan Lantz at 817-200-1173 and in writing within 10 days from the issue date of
this letter to notify the NRC of your intentions. If we have not heard from you within 10 days, we
will continue with our significance determination and enforcement decision. The final resolution
of this matter will be conveyed in separate correspondence.
Because the NRC has not made a final determination in this matter, no Notice of Violation is
being issued for these inspection findings at this time. In addition, please be advised that the
number and characterization of the apparent violation described in the enclosed inspection
report may change as a result of further NRC review.
In accordance with 10 CFR 2.390 of the NRC's "Rules of Practice," a copy of this letter and its
enclosure will be made available electronically for public inspection in the NRC Public
P. Dietrich
- 4 -
Document Room or from the NRCs document system (ADAMS), accessible from the NRC Web
site at http://www.nrc.gov/reading-rm/adams.html (the Public Electronic Reading Room).
Sincerely,
/RA/
Steven A. Reynolds
Acting Regional Administrator
Dockets: 50-361, 50-362
Enclosure:
NRC Inspection Report 05000361/2012009
and 05000362/2012009
Attachments:
1. Supplemental Information
2. Independent Evaluation of San Onofre
Nuclear Generating Station (SONGS)
Steam Generator Tube Wear Problems
3. NRC International Travel Trip Report
4. Report to NRC, Submitted by V.K. Dhir
5. Preliminary Significance Determination
Loss of Steam Generator Tube Integrity
cc w/enclosure:
Electronic Distribution
ADAMS: No Yes
SUNSI Review Complete
Reviewer Initials: GEW
Publicly Available
Nonsensitive
Nonpublicly Available
Sensitive
RIV:RI:SPB
RI:SPB
SSP:NRR
SSP:NRO
SRA:DRS:EB1 I&AL:SPB:ORA RSLO:ORA
MRBloodgood
JPReynoso
ELMurphy
CGThurstson GDReplogle
GEWerner
WAMaier
E - GEWerner E - GEWerner E - GEWerner E - GEWerner /RA/
/RA/
/RA/
6/28/13
7/1/13
6/29/13
7/12/13
7/27/13
7/12/13
7/15/13
SPAO:ORA
SES:ACES
C:ACES/ORA
C:SPB/ORA
RC/ORA
TM:SSP
Acting RA
VLDricks
RSBrowder
HJGepford
RELantz
KDFuller
ATHowell
SAReynolds
/RA/
/RA/
/RA/
E - GEWerner /RA/
E - GEWerner /RA/
7/16/13
7/23/13
8/23/13
8/29/13
9/4/13
9/20/13
9/20/13
Enclosure
TABLE OF CONTENTS
SUMMARY OF FINDINGS ........................................................................................................... 3
Summary of Plant Status .............................................................................................................. 5
1.
REACTOR SAFETY .............................................................................................................. 5
4OA3 Follow-up of Events and Notices of Enforcement Discretion (71153) ............................... 5
.1
Event Report Review ............................................................................................................ 5
4OA5 Other Activities .................................................................................................................. 6
.1
Review of Root Cause Evaluations .................................................................................. 6
.2
Operational Assessments ............................................................................................... 12
.3
Thermal-Hydraulic and Vibration Models ........................................................................ 14
.4
Design Modification Review ............................................................................................ 16
.5
(Closed) Unresolved Item 05000362/2012007-08, Non-Conservative Thermal-Hydraulic
Model Results ................................................................................................................ 20
.6
(Closed)05000362/2012007-04 Evaluation of Changes in Dimensional Controls during
the Fabrication of Unit 2 and Unit 3 Replacement Steam Generators .......................... 29
.7
Resolution of Independent Technical Review Findings.31
.8 Chalk River Testing.....40
4OA6 Meetings ......................................................................................................................... 41
Exit Meeting Summary ........................................................................................................... 41
SUPPLEMENTAL INFORMATION .......................................................................................... A1-1
LIST OF ITEMS OPENED, CLOSED, AND DISCUSSED .................................................. A1-1
LIST OF DOCUMENTS REVIEWED .................................................................................. A1-2
Independent Evaluation Of San Onofre Nuclear Generating Station (SONGS) Steam
Generator Tube Wear Problems ......................................................................................... A2-1
NRC International Travel Trip Report ..A3-1
Report to NRC, Submitted by V.K. Dhir.......A4-1
Preliminary Significance Determination of Loss of Steam Generator Tube IntegrityA5-1
- 2 -
U.S. NUCLEAR REGULATORY COMMISSION
REGION IV
Docket:
05000361, 05000362
License:
Report:
05000361/2012009 and 05000362/2012009
Licensee:
Southern California Edison Company
Facility:
San Onofre Nuclear Generating Station, Units 2 and 3
Location:
5000 S. Pacific Coast Highway
San Clemente, California
Dates:
December 3, 2012, through June 7, 2013
Team Lead:
G. Werner, RIV, SONGS Project Branch, Inspection Lead
Inspectors:
R. Lantz, Chief, SONGS Project Branch
J. Reynoso, RIV, Resident Inspector
E. Murphy, NRR, Senior Materials Engineer
C. Thurston, RES, Reactor Systems Engineer
M. Bloodgood, RIV, Reactor Engineer
Accompanying
Personnel:
J. Billoue, Steam Generator Design Engineer, Beckman and Associates
V. Dhir, PhD, Thermal-Hydraulic and Vibration Contractor, Beckman and
Associates
G. Warrier, PhD, Laboratory Research Engineer Contractor, Beckman and
Associates
Approved By:
A. Howell, Team Manager
San Onofre Nuclear Generating Station Special Project
- 3 -
SUMMARY OF FINDINGS
IR 05000361/2012009; 05000362/2012009; 12/03/2012 - 06/07/2013; San Onofre Nuclear
Generating Station; Confirmatory Action Letter Response Inspection.
This inspection team was comprised of one resident, two region-based, two headquarters-based,
and three contractor inspectors. One apparent White violation of low to moderate safety
significance and one Green noncited violation of very low safety significance were identified. The
significance of most findings is indicated by their color (Green, White, Yellow, or Red) using
Inspection Manual Chapter 0609, Significance Determination Process. The cross-cutting
aspect is determined using Inspection Manual Chapter 0310, Components Within the Cross-
Cutting Areas. Findings for which the significance determination process does not apply may be
Green or be assigned a severity level after NRC management review. The NRCs program for
overseeing the safe operation of commercial nuclear power reactors is described in NUREG-
1649, Reactor Oversight Process, Revision 4, dated December 2006.
A. NRC-Identified Findings and Self-Revealing Findings
Cornerstone: Initiating Events
Green. The inspectors identified a noncited violation of 10 CFR Part 50, Appendix B,
Criterion III, Design Control, for the failure to verify the adequacy of the thermal-hydraulic
and flow-induced vibration design of the Unit 2 replacement steam generators, resulting in
excessive and unexpected steam generator tube wear after one cycle of operation. The
licensee initiated Nuclear Notification NN 202447268 to address this issue in the corrective
action program. Southern California Edison revised the thermal-hydraulic code of record and
ensured that the code was in accordance with ASME guidance. Subsequently, on June 7,
2013, Southern California Edison announced that Units 2 and 3 would be permanently shut
down.
The finding is more than minor because it is associated with the equipment performance
attribute of the Initiating Event Cornerstone and adversely affected the cornerstone objective
of limiting the likelihood of those events that upset plant stability and challenge critical safety
functions during shutdown as well as power operations. The inspectors used NRC
Inspection Manual Chapter 0609, Attachment 4 and Appendix A, to evaluate the significance
of this finding. In accordance with Exhibit 1 of Inspection Manual Chapter 0609, Appendix A,
the inspectors determined that the finding was of very low safety significance because the
finding did not involve a degraded steam generator tube that could not sustain three times
the normal operating differential pressure and did not violate the accident leakage
performance criterion. No cross-cutting aspect was assigned because this performance
deficiency occurred in the 2005 to 2008 timeframe. Substantial management and personnel
changes have occurred, including taking actions to address a chilled work environment and
other safety culture issues. The NRC determined that the performance behavior that existed
at that time is not indicative of current performance.
- 4 -
Apparent Violation. The inspectors identified an apparent violation of 10 CFR Part 50,
Appendix B, Criterion III, Design Control, for the failure to verify the adequacy of the
thermal-hydraulic and flow-induced vibration design of the Unit 3 replacement steam
generators, which resulted in significant and unexpected steam generator tube wear after
11 months of operation and an associated apparent violation of Technical Specification
5.5.2.11, Steam Generator Program, loss of tube integrity on Unit 3 Steam Generator 3E0-
88. The licensee initiated Nuclear Notification NN 202447265 to address this issue in the
corrective action program. Southern California Edison revised the thermal-hydraulic code of
record and ensured that the code was in accordance with ASME guidance. Subsequently, on
June 7, 2013, Southern California Edison announced that Units 2 and 3 would be
permanently shut down.
This finding is more than minor because it is associated with the equipment performance
attribute of the Initiating Events Cornerstone and adversely affected the cornerstone objective
of limiting the likelihood of those events that upset plant stability and challenge critical safety
functions during shutdown as well as power operations. Specifically, the failure to verify the
adequacy of the thermal-hydraulic and flow-induced vibration design resulted in excessive
and rapid tube wear due to fluid elastic instability, which challenged the structural integrity of
the tubes to perform their pressure boundary function. The inspectors used NRC Inspection
Manual Chapter 0609, Attachment 4 and Appendix A, to evaluate the significance of this
finding. In accordance with Exhibit 1 of Inspection Manual Chapter 0609, Appendix A, the
inspectors determined that this finding required evaluation in accordance with Inspection
Manual Chapter 0609, Appendix J, because the finding involved a degraded steam generator
tube condition where one tube could not sustain three times the differential pressure across a
tube during normal full power, steady-state operation. In accordance with Inspection Manual
Chapter 0609, Appendix J, this finding required a detailed risk analysis, since it involved two
or more tubes that could not sustain three times the normal differential pressure and one or
more steam generators that violated accident-induced leakage performance criterion. A
Phase 3 analysis was completed using the San Onofre SPAR model, Revision 8.22,
assuming average test and maintenance, and a truncation limit of 1.0E-11. Based on the
best available information, the performance deficiency was preliminarily characterized as a
finding of low to moderate safety significance (White). The final significance of this finding is
to be determined. No cross-cutting aspect was assigned because this performance
deficiency occurred in the 2005 to 2008 timeframe. Substantial management and personnel
changes have occurred, including taking actions to address a chilled work environment and
other safety culture issues. The NRC determined that the performance behavior that existed
at that time is not indicative of current performance
B. Licensee-Identified Violations
None.
- 5 -
REPORT DETAILS
Summary of Plant Status
Prior to the Unit 3 steam generator tube leak, Unit 2 was shut down for a scheduled refueling
outage and Unit 3 was operating at 100 percent rated thermal power with no plant evolutions in
progress. On January 31, 2012, Unit 3 control room operators received an alarm that indicated a
primary-to-secondary reactor coolant leak from Steam Generator 3E0-88. The alarm received
was from the main condenser air ejector radiation monitors, which continuously sample from a
vent line for the purpose of rapidly identifying steam generator tube leaks. Although the leak rate
was small, it increased enough in a short period of time for the licensee to perform a rapid
shutdown. The estimated leak rate was 75 gallons per day. The facility license allows full power
operation with a steady-state leak rate of less than 150 gallons per day. On February 2, 2012,
Unit 3 reached cold shutdown conditions. The licensee reviewed the amount of gaseous
radioactivity released and estimated a dose of approximately 0.0000452 mrem to a member of
the public. The annual regulatory limit to a member of the public is 100 mrem per year. At the
time of the inspection, Units 2 and 3 continued to remain in a shutdown status, with Unit 3
defueled.
On June 7, 2013, Southern California Edison (SCE) announced that Units 2 and 3 would be
permanently shut down. On June 12, SCE submitted a Certification of Permanent Cessation of
Power Operations to the NRC, certifying that Units 2 and 3 have permanently ceased power
operations. On June 28 and July 22, 2013, SCE certified that all fuel had been permanently
removed from the Units 3 and 2 reactors, respectively (ML13183A391 and ML13204A304).
1. REACTOR SAFETY
4OA3 Follow-up of Events and Notices of Enforcement Discretion (71153)
.1
Event Report Review
a. Inspection Scope
The inspectors reviewed the following Licensee Event Report and related documents to
assess: (1) the accuracy of the Licensee Event Report; (2) the appropriateness of
corrective actions; (3) violations of requirements; and (4) generic issues.
(Closed) Licensee Event Report 05000362/2012-002-00, Unit 3 Steam Generator Tube
Degradation Indicated by Failed In-Situ Pressure Testing
As described above, Unit 3 was shut down on January 31, 2012, because of a small
steam generator tube leak. Subsequent in-situ pressure testing was conducted in March
2012 on a total of 129 tubes in Steam Generators 3E0-88 and 3E0-89. Eight tubes failed
in-situ pressure testing in Steam Generator 3E0-88, with no tube failures in Steam
Generator 3E0-89.
- 6 -
The inspectors identified an apparent violation of 10 CFR Part 50, Appendix B,
Criterion III, Design Control, for not verifying the design of the replacement steam
generators thermal-hydraulic and vibration analysis, which resulted in multiple losses of
tube integrity contrary to Technical Specification 5.5.2.11.
See Report Section 4OA5, Subsection 5, for a detailed description of the closure of the
unresolved item associated with the Licensee Event Report.
4OA5 Other Activities
Inspection Procedure 92702, Followup on Traditional Enforcement Actions Including
Violations, Deviations, Confirmatory Action Letters, Confirmatory Orders, and Alternative
Dispute Resolution Confirmatory Orders
.1
Review of Root Cause Evaluations
a. Inspection Scope
The inspectors reviewed SCE Root Cause Evaluation Nuclear Notifications
201836126 and 201836127 and the Mitsubishi Heavy Industries (Mitsubishi)
Technical Evaluation Report, Document L5-04GA564, to determine the following:
(1) complete and accurate identification of the problem; (2) evaluation and disposition
of operability/reportability issues; (3) consideration of extent of condition, generic
implications, and common cause; (4) classification and prioritization of the resolution
of the problem; (5) identification of root and contributing causes of the problem;
(6) identification of corrective actions; (7) completion of corrective actions on Unit 2,
including corrective actions based on the Unit 3 cause evaluation; and
(8) effectiveness reviews.
b. Observations and Findings
No findings were identified.
The inspectors reviewed the SCE root cause evaluations and Mitsubishi technical
evaluation related to the SCE steam generator tube degradation issue. The licensee
and Mitsubishi used multiple techniques, which included event and causal factor
charting, barrier analysis, and Kepner-Tregoe analysis to assess the probable causes
of the tube wear identified in both Units 2 and 3 steam generators. Each steam
generator has 9727 U-tubes, which are supported by seven tube support plates and
six sets of V-shaped anti-vibration bars. The inspectors identified that both SCE and
Mitsubishi identified four types of wear in both Units 2 and 3 steam generators. The
identified wear was assessed by both SCE and Mitsubishi and corrective action
recommendations were identified. The four types of tube wear are discussed in the
following sections of this report.
- 7 -
Type 1 (tube-to-tube wear) - Wear in the tube free-span sections in the U-bend
region. Most of the tubes with this type of wear also have wear indications at anti-
vibration bars and tube support plates. In this case, it is considered that the entire
tube, including the straight leg, was vibrating excessively.
Type 2 (anti-vibration bar wear) - Wear at only the tube-to-anti-vibration bar
intersections, with no wear indications in the tube free-span sections. Some of
these tubes have wear indications at the tube support plates as well. In this case,
it is considered that mainly the U-bend section of the tube was vibrating.
Type 3 (tube support plate wear) - Wear at the tube-to-tube support plate
intersections only in the straight section of the tubes. In this case, it is considered
that only the straight section of the tube was vibrating.
Type 4 (retainer bar wear) - Wear at the anti-vibration bar structure retainer bars
in the tube U-bend section. These tubes have no wear indications in the free
span, at anti-vibration bars or at tube support plates. In this case, it is considered
that the retainer bar itself was vibrating and the tube was not vibrating.
(1)
Tube-to-Tube Wear (Type 1)
The inspectors reviewed the licensees extent of condition associated with the
identified tube-to-tube wear, which affected 326 tubes in Unit 3 steam generators
during eddy current testing following the steam generator leak identified on
January 31, 2012. Unit 2 was in a current shutdown for the first refueling outage
following the replacement of both of the steam generators. The licensee had
completed an initial 100 percent bobbin coil eddy current inspection of the Unit 2
steam generators, which failed to identify any tube-to-tube wear. Following the
identification of the tube-to-tube wear in the Unit 3 steam generators, the licensee
performed eddy current testing using a more sensitive (+P) probe of
approximately 1300 tubes in each of the Unit 2 steam generators from the same
area as the identified Unit 3 tube-to-tube wear. The licensee identified two tubes
with approximately 15 percent tube-to-tube wear that were not identified during
the initial eddy current testing using the bobbin coil probe.
The inspectors reviewed the mechanisms that were determined to be contributors
to the tube-to-tube wear by the licensee and Mitsubishi. The mechanical cause of
the Unit 3 tube-to-tube wear was determined by SCE and Mitsubishi to be fluid-
elastic instability associated with adverse secondary thermal-hydraulic conditions
and lack of effective in-plane tube support for the tubes. Mitsubishi determined
that the tube-to-anti-vibration bar contact forces used in the replacement steam
generators was insufficient to prevent the in-plane motion given the thermal-
hydraulic conditions in the secondary side of the steam generators. Mitsubishi
identified, as part of their evaluation, that the contact forces in Unit 3 were less
than the Unit 2 contact forces following the review of the manufacturing
dimensional tolerances. The anti-vibration bar dimensional tolerances are further
discussed in Section 6 of this report. Westinghouses independent assessment
- 8 -
(SG-SGMP-12-10, Operational Assessment of Wear Indications in the U-bend
Region of San Onofre Nuclear Generating Station Unit 2 Replacement Steam
Generators Supporting Restart, Revision 3) concluded that Unit 2s tube-to-tube
wear may be a result of tube-to-tube proximity in conjunction with flow-induced
vibration. This mechanism was determined by Westinghouse to be a probable
cause due to the wear patterns on the anti-vibration bars being limited to the anti-
vibration bar width and not exhibiting the longer wear patterns associated with in-
plane movement. The inspectors identified that this extent of condition for the
tube proximity wear mechanics was not evaluated as part of the root cause
evaluation. The licensee issued Nuclear Notification NN 201836127 to update the
tube-to-tube wear root cause analysis with the licensees and Mitsubishis analysis
of the Westinghouse Operations Assessment manufacturing issues and any
resulting corrective action to address these issues.
The inspectors reviewed the licensees corrective actions associated with the
tube-to-tube wear. The licensee initially plugged the two tubes that were identified
as having tube-to-tube wear in Unit 2. In addition, the licensee preventively
plugged 321 Unit 2 tubes using selective process information from the Unit 3
steam generator wear data. The preventive tube plugging selection processes
used nine screening criteria, including the location of anti-vibration bars and tube
support plate wear indications, length of anti-vibration bar wear indications,
average void fraction over the length of the tube, location of the tube within the
bundle, and coupling between adjacent susceptible tubes. Results of the
assessments of each tube against the nine screening criteria were reviewed
cumulatively to identify which tubes would be preventively plugged. No issues
were identified.
(2)
Tube to Anti-Vibration Bar Wear (Type 2)
The inspectors reviewed the licensees extent of condition associated with the
identified tube-to-anti-vibration bar wear, which affected 1767 tubes in Unit 3 and
1399 tubes in Unit 2 steam generators. The licensee considered tube-to-anti-
vibration bar wear patterns as part of their evaluation consisting of: (1) wear
patterns in tubes that exhibited tube-to-tube wear and (2) wear patterns in tubes
that did not exhibit tube-to-tube wear. The first pattern was determined to be the
result of conditions resulting in fluid-elastic instabilities and subsequent in-plane
motion of the tubes. These tube-to-anti-vibration bar wear patterns tended to
extend beyond the edges of the anti-vibration bar, indicating an in-plane sliding
motion of the U-bend that also led to the tube-to-tube wear in the affected tube.
The second tube-to-anti-vibration bar wear pattern was considered to be due to
turbulence-induced vibration in the out-of-plane direction. The length of these
wear patterns was confined to the width dimension of the anti-vibration bars.
Mitsubishi only discussed the second wear pattern described by SCE as part of
their Type 2 wear pattern. The first wear pattern was addressed as part of the
tube-to-tube wear (Type 1) discussion. Mitsubishi determined that the wear was
due to random tube vibration. Mitsubishi describes random vibration as a
phenomenon where the tubes vibrate due to forces created by turbulent flow as a
- 9 -
result of fluid velocity and density fluctuations, which are smaller than those due to
tube fluid-elastic instability. Mitsubishis determination was consistent with SCE
review of the second wear pattern. The adverse secondary thermal-hydraulic
condition and the lack of effective anti-vibration bar supports contributed to the
presence of fluid-elastic instabilities and turbulence induced vibration. The
licensee stabilized and plugged four of these tubes in Unit 2 and one tube in
Unit 3 in accordance with their Steam Generator Program due to the wear
identified during the eddy current inspection.
(3)
Tube-to-Tube Support Plate Wear (Type 3)
The inspectors reviewed the licensees extent of condition associated with the
identified tube-to-tube support plate wear, which affected 463 tubes in Unit 3 and
299 tubes in Unit 2 steam generators. The licensee considered two categories of
wear as part of their evaluation consisting of: (1) tube-to-tube support plate wear
affecting tubes also exhibiting tube-to-tube wear, and (2) tube-to-tube support
plate wear in tubes not exhibiting tube-to-tube wear. The licensee determined that
the higher tube-to-tube support plate wear identified in the Unit 3 steam generator
was due to in-plane fluid-elastic instability, resulting in higher displacement
vibrations. This conclusion was based on the relationship between the identified
tubes with tube-to-tube support plate and tube-to-tube wear. Mitsubishi
considered Type 3 wear as only straight leg wear due to vibrations corresponding
to the second category described by SCE. Mitsubishi concluded that the wear
was caused by cross-flow induced random vibration in the region where
secondary fluid cross-flow velocities are high. The licensee did not perform
additional cause analysis for the tube-to-tube support plate wear due to the close
correlation with the tube-to-tube wear and the corrective actions being inclusive in
the tube-to-tube wear corrective actions.
(4)
Tube-to-Retainer Bar Wear (Type 4)
The licensee requested Mitsubishi to conduct an evaluation of the tube-to-retainer
bar wear as a result of the wear indications found in the Unit 2 steam generators.
Mitsubishis evaluation determined that the wear was the result of movement of
the retainer bar located at anti-vibration bars 2 and 3 on the hot leg side of the U-
bend and anti-vibration bars 10 and 11 on the cold leg side. A review of the
retainer bars at these locations revealed that the replacement steam generator
retainer bars were significantly longer (24.02 inches in length) when compared to
other steam generator designs (7-13 inches in length). In addition, the natural
frequency associated with the longer, thinner retainer bars was significantly
(approximately 5 times) less than the other steam generator designs reviewed
during the evaluation. The licensee determined that the mechanical root cause of
the wear was the combination of the longer length and smaller diameter (0.187
inches) retainer bar natural frequency and the increased flow velocities, which
contributed to a flow-induced vibration of the retainer bars, resulting in tube-to-
retainer bar contact.
- 10 -
The licensee determined that the lack of tube-to-retainer bar wear at anti-vibration
bars 1 and 12, which are the same dimensions as the four previously discussed,
was attributed to the lower flow velocity (7.5 ft/s) at the anti-vibration bar 1 location
compared to the flow velocity (9.8 ft/s) at the anti-vibration bar 2 location.
Mitsubishi failed to evaluate the potential effects of flow-induced vibrations during
the anti-vibration and retainer bar design, due to assumptions that the natural
frequency of the retaining bar was high enough to preclude flow-induced vibration.
Mitsubishi failed to consider the effects of using retainer bars that were longer and
of smaller diameter than those they had previously used.
During the design of the replacement steam generators, SCE questioned
Mitsubishi on the lack of vibration analysis of the retainer bars but failed to
independently evaluate the Mitsubishi response and agreed with the
determination that the retainer bars would not come into contact with the tubes.
Mitsubishi provided additional chromium plating of the retainer bars to reduce the
wear coefficient and minimize potential wear in response to the licensees
questions.
The licensees corrective actions included plugging 94 tubes in each of the Units 2
and 3 steam generators in the vicinity of the retainer. The licensee also stabilized
14 tubes in the Unit 2 and 12 tubes in the Unit 3 steam generators. For details
about the retainer bar design noncited violation, please refer to NRC Inspection
Report 05000361/2012010 and 05000362/2012010 (ML12318A342).
(5)
Return to Service Defense-in-Depth Actions
The inspectors reviewed the licensees defense-in-depth actions, which were
specified in Section 9 of the San Onofre Nuclear Generating Station Unit 2 Return
to Service Report dated October 3, 2012. The inspectors reviewed the following
defense-in-depth actions and determined them to be enhancements which assist
the operators identification and response in the event a steam generator tube
leak occurs:
- 11 -
(a) Injection of Argon into the Reactor Coolant System
The inspectors reviewed the defense-in-depth action of adding argon to the
reactor coolant system. The addition of argon was planned to be added to
enhance/improve the primary-to-secondary leak rate detection level at the
condenser air ejector radiation monitor. The inspectors reviewed the
licensees actions to add argon to improve the leak rate sensitivity. The
inspectors reviewed Procedure SO123-III-2.22.23, Unit 2/3 Steam Generator
Tube Leakage Monitoring Program, Revision 25, which specified that the
reactor coolant system activated argon activity should be maintained between
0.05 uCi/ml and 0.15 uCi/ml. A value of 0.10 uCi/ml of activated argon
ensures that the condenser air ejector radiation monitor is capable of detecting
a 5 gpd primary-to-secondary leak rate instead of the normal setpoint of
30 gpd. In addition, the inspectors reviewed argon injection and controls
specified in Procedures SO23-3.2.1, CVCS Operation, Revision 40, and
SO23-3.2.1.1, CVCS Alignment, Revision 20.
(b) Installation of Nitrogen-16 Radiation Detection System on the Main Steam
Lines
The inspectors reviewed the licensees implementation of a radiation detection
system for detecting nitrogen-16. The licensee installed nitrogen-16 detectors,
a more sensitive radiation detection system, adjacent to the main steam line to
provide early operational responses to primary-to-secondary leaks in the
steam generators in addition to current radiation detection systems. The
nitrogen-16 detectors are located so that a leak would be detected seconds
after it occurs. This is an improvement from the current condenser offgas
radiation monitors and blowdown samples, which could take up to one hour.
The inspectors reviewed Nuclear Notification NN 800905312, which described
the implementation of the nitrogen-16 detectors and procedural guidance for
the operation of the detecting system. The inspectors identified that
Procedure SO23-3-2.24, Radiation Monitoring System Guidelines and RDU
Operation, Revision 14, was changed to direct actions to implement
Procedure SO23-13-14, Reactor Coolant Leak, Revision 21, primary-to-
secondary operator actions in the event that the nitrogen-16 monitor alarms.
This will provide for earlier operator actions in the event of a primary-to-
secondary leak.
(c) Reduction of Administrative Limits for the Reactor Coolant System Activity
Level
The inspectors reviewed changes to Procedure SO123-III-1.1.23, Unit 2/3
Chemistry Control of Primary Plant and Related Systems, Revision 60, for
defense-in-depth actions associated with administrative limits for reactor
coolant system activity levels. The inspectors identified that Nuclear
Notification NN 201836127, Task 39, lowered the dose equivalent iodine-131
normal range in Procedure SO123-III-1.1.23 from 1.0 µCi/gram to
- 12 -
0.5 µCi/gram. This provides an additional action level prior to reaching the
technical specification limit of 1.0 µCi/gram. The inspectors determined that
procedural guidance will require Operations personnel to make a
determination of continued plant operation if the dose equivalent iodine-131
exceeds the new lower normal range (0.5 µCi/gram) instead of the technical
specification limit.
(d) Enhanced Operator Response to Early Indication of Steam Generator Tube
Leakage
The inspectors reviewed procedural changes, training material, and training
records, in addition to interviewing licensee personnel associated with the
enhanced operator response. The training consisted of classroom lectures;
simulator scenarios; and just-in-time training associated with steam generator
tube ruptures, the addition of argon and temporary nitrogen-16 monitors, and
changes in plant procedures related to steam generator tube ruptures, primary
leaks, and modifications. This training was performed as part of the operator
requalification program. The inspectors reviewed the training completion
records associated with the training related to the enhanced operator actions
and identified that the training was being performed but that not all operators
had completed all of the training at the time of the inspection. The licensee is
tracking the training completion in their operator requalification program.
The inspectors determined that the defense-in-depth actions were determined to
be appropriate and could be performed in accordance with procedural guidance.
.2
Operational Assessments
a. Inspection Scope
The inspectors reviewed the SCE operational assessment of steam generator tube
integrity for Unit 2 for the period extending from Unit 2 restart from Refueling
Outage 17 to Unit 2 shutdown for its next scheduled steam generator inspection. The
inspectors reviewed the following specific documents pertaining to the operational
assessment:
SCE Confirmatory Action Letter response dated October 3, 2012, Enclosure 2,
San Onofre Nuclear Generating Station Unit 2 Return to Service Report
San Onofre Nuclear Generating Station Unit 2 Return to Service Report,
Attachment 4, MHI Document L5-04GA564, Tube Wear of Unit-3 RSG - Technical
Evaluation Report, Revision 9 (proprietary version), prepared by Mitsubishi
San Onofre Nuclear Generating Station Unit 2 Return to Service Report,
Attachment 6, SONGS U2C17 Steam Generator Operational Assessment:
- 13 -
o Appendix A, Document 1814-AU651-MO144, SONGS U2C17 Outage -
Steam Generator Operational Assessment, Revision 0 (proprietary version),
prepared by AREVA NP Inc. for degradation mechanisms other than tube-to-
tube wear
o Appendix B, Document 1814-AU651-MO146, SONGS U2C17 Steam
Generator Operational Assessment for Tube-to-Tube Wear, Revision 0
(proprietary version), prepared by AREVA NP Inc.
o Appendix C, Document 1814-AU651-MO145, Operational Assessment for
SONGS Unit 2 SG for Upper Bundle Tube-to-Tube Wear Degradation at the
End of Cycle 16, Revision 1, prepared by Intertec APTECH
o Appendix D, Document 1814-AA086-M0190, Operational Assessment of
Wear Indications in the U-Bend Region of San Onofre Nuclear Generating
Station Unit 2 Replacement Steam Generators, Revision 4, prepared by
Westinghouse Electric Company, LLC.
The inspectors assessed the implementation of the operational assessments relative
to Technical Specification 5.5.2.11, Steam Generator (SG) Program, and Electric
Power Research Institute (EPRI) Report 1019038, Steam Generator Management
Program: Steam Generator Integrity Assessment Guidelines, Revision 3, as
referenced in Procedure SO23-SG-1, SONGS Steam Generator Program,
Revision 20. The inspectors review included the following items:
Degradation mechanisms, growth rate calculations, and assumptions
Tube plugging and stabilization
Operating restrictions
Appropriateness of 5-month inspection interval
Application of Unit 3 extent-of-condition on Unit 2 operational assessments
Midcycle inspection methodology
The inspectors also reviewed the thermal-hydraulic models and flow-induced vibration
models used for the operational assessments. This part of the review is addressed in
Section 4OA5.3 of this inspection report.
b. Observations and Findings
Based on SCEs decision to retire both units, the reviews of the operational
assessments were not completed. No conclusions were made as to the adequacy of
each operational assessment.
- 14 -
.3
Thermal-Hydraulic and Vibration Models
a. Inspection Scope
The inspectors reviewed the Mitsubishi ATHOS thermal-hydraulic model as well as
the FIVATS tube vibration models as specified in Specification SO23-617-1,
Specification for Design and Fabrication of the Replacement Steam Generators for
Unit 2 and Unit 3, Revision 4. As part of the review, the inspectors compared the
results of the Mitsubishi ATHOS thermal-hydraulic model results to results of AREVA
CAFCA4 and Westinghouse ATHOS. The NRC independently ran the ATHOS model
to verify consistency and appropriateness of the steam velocities and void fractions.
The inspectors also had the assistance of a contractor with expertise in steam
generator tube vibration who reviewed the technical basis for the revised Connors
equation used to calculate stability ratios in the in-plane direction.
b. Observations and Findings
No findings were identified.
ATHOS is an industry three-dimensional computational fluid dynamics code
developed by EPRI to assess thermal-hydraulic conditions in steam generators. The
code was developed in the 1980s and is still being used today by nuclear facilities
and vendors. The code iteratively solves the conservation of the mass, momentum,
and energy equations along with empirical correlations to calculate the thermal-
hydraulic parameters for a given steam generator geometry design and set of plant
operating conditions. The code has many other uses, including performance
trending, deposit mapping, and sludge pile predictions. The code is not used for
safety analysis and has not been reviewed or approved by the NRC.
Mitsubishi used the latest version, EPRI ATHOS version 3.1, of the code while
Westinghouse used its ATHOS60 version 3.0. Westinghouse has the most extensive
experience with the code, having developed several pre- and post-processor add-ons
for its analysis methodology.
Westinghouse uses the ATHOS code for new designs as well as assessments for the
various model series of Westinghouse steam generators. Westinghouse performed
validation and verification of ATHOS by benchmarking the code against several full-
sized and scaled model tests1. They concluded that the ATHOS calculated thermal-
hydraulic parameters were in good agreement with measured test data.
Mitsubishi and Westinghouse each independently developed ATHOS models and ran
cases to generate three-dimensional thermal-hydraulic performance data for various
specific SCE steam generator boundary conditions. The NRC also developed an
ATHOS model and ran limited cases for independent assessment of the vendor
analyses.
1 Westinghouse LTR-SGDA-12-50 dated 10/14/2012
- 15 -
Both Mitsubishi (L5-04GA566) and Westinghouse (LTR-SMP-12-36) ran cases on
Units 2 and 3 for plugging scenarios and for power levels ranging from
50-100 percent in 10 percent increments. Velocity and density output from these
cases were used as input to the flow induced vibration analysis. For the vibration
analysis, Westinghouse uses FASTVIB and Mitsubishi uses FIVATS. Both analyses
compute the stability ratios based on a form of the Connors Equation, but there were
considerable differences in the empirical constants used to characterize the tube
excitation threshold. The empirical constants are the critical factor K and the overall
tube damping ratio h. The FASTVIB code incorporates the analytical approaches with
constants that were largely defined by the work of H. J. Connors while conducting
research for Westinghouse at their research lab. The analytical approach of
Mitsubishi was a best estimate and was principally based on more recent work by the
Canadian researcher, M.J. Pettigrew (École Polytechnique de Montreal).
Prevention of excessive vibration and fretting wear is generally achieved by a
combination of design, analysis, and testing as defined by each vendor and their own
methodology. Each vendor methodology is required to have some level of validation,
as noted for design of ASME components. This methodology is used to lay out and
evaluate the many aspects of design before the steam generators are manufactured.
The anti-vibration bars and tube support plates should be arranged to meet specified
design limits established to prevent fluid instabilities and minimize tube wear. The risk
of vibration is highest in the U-bend region, where velocity and void fraction are
highest or there is significant unsupported span length which lowers the natural
frequency of the tube making that span more susceptible to vibration.
The methodology, based on the Connors Equation, is of the following general form:
.
(1)
where:
Uc - Critical flow velocity
K - Critical factor
h - damping ratio
- fluid density surrounding the tube
and where:
(2)
Ueff - effective gap velocity (computed from ATHOS)
SR - stability ratio
The computations for K and h vary for in- and out-of-plane stability ratio and are
different in each vendors methodology. The vibration analysis methods used by
Mitsubishi were similar to that recommended by ASME Code Section III, Appendix N,
originally designed for traditional out-of-plane stability analysis. The anti-vibration
bars were designed primarily to prevent out-of-plane motion. The design thickness of
the anti-vibration bars provides for a small gap that essentially closes when the plant
- 16 -
heats up to normal operating temperature. The damping improves as the gaps close,
providing stability for the tube in the out-of-plane direction. Although not specifically
part of the design, anti-vibration bars are known to provide some limited stability and
support for tube motion in the in-plane direction based on experimental and empirical
information.
Historically, during the design of all steam generators, including the SCE replacement
steam generators, in-plane vibration was not considered since it had not been seen in
any operating steam generators and was believed bounded by (i.e., can occur only as
a consequence of) out-of-plane vibration analysis.
Based on the steam generator tube-to-tube wear caused by in-plane motion, both
vendors modified the Connors Equation for applicability to in-plane fluid-elastic
instability based on a mixture of internal experimental methods and academic
research. The methodology is highly dependent upon the number of assumed
continuous ineffective anti-vibration bars and other assumptions that have not been
verified by experimental data.
In licensee calculations at 100 percent power with all anti-vibration bar supports
effective, neither of the vendor models predicted stability ratios above 1.0. Both the
Mitsubishi and Westinghouse analyses have evaluated a minimum of two continuous
ineffective anti-vibration bars, and the calculations then advance in continuous
increments of two ineffective supports until in-plane fluid-elastic instability is predicted,
i.e., stability ratio greater than 1.0 with various levels of plant power.
Overall, the inspectors determined that the Mitsubishi vibration stability ratio results
were higher than Westinghouses independent analysis and, therefore, provided some
confirmation that Mitsubishis model was not under-predicting the conditions of steam
generators.
The NRC did not develop independent vibration calculations; however, the vendor
computations for vibration were independently reviewed and checked for correctness.
The inspectors did not identify any issues with the vibration calculations.
.4
Design Modification Review
a. Inspection Scope
The inspectors reviewed the design changes, listed below, that were associated with
the licensees Confirmatory Action Letter Response for Unit 2 to determine whether
the changes to the facility or procedures, as described in the Updated Final Safety
Analysis Report, had been reviewed and documented in accordance with 10 CFR
50.59 requirements.
Plant operation at 70 percent power level
Argon (Ar-40) injection into the reactor coolant system
Nitrogen (N-16) radiation detection system on the main steam lines
- 17 -
Vibration and loose parts monitoring system
Annunciator for 70 percent power operation
Tube plugging and stabilization
The inspectors reviewed the various information used by SCE to make the changes to
Unit 2 associated with their return to operation, including calculations, analyses,
design change documentation, procedures, the Updated Final Safety Analysis Report,
the Technical Specifications, and plant drawings. The inspectors interviewed plant
personnel responsible for developing and evaluating the design changes. The
inspectors compared the safety evaluations and supporting documents to the
guidance and methods provided in NEI 96-07, "Guidelines for 10 CFR 50.59
implementation," Revision 1, as endorsed by NRC Regulatory Guide 1.187,
"Guidance for Implementation of 10 CFR 50.59, Changes, Tests, and Experiments," to
determine the adequacy of the evaluations.
b. Observations and Findings
The inspectors determined that the 50.59 screen and evaluations were performed in
accordance with the requirements of 10 CFR 50.59, Changes, Tests, and
Experiments," with the exception of the following discussions.
During the inspectors review of the 10 CFR 50.59 screenings for operation at
70 percent reactor power associated with Nuclear Engineering Change Package
800873488-0131 and vibration and loose parts monitoring system modification
associated with Nuclear Engineering Change Package 800457837-0550, the
following deficiencies were identified:
(1)
Operation at 70 percent Reactor Power
(a) Technical Specification Surveillance Requirement 3.3.1.11 states, Using
the incore detectors, verify the shape annealing matrix elements to be used
by the CPCs [core protection calculators], with a specified frequency of
Once after each refueling prior to exceeding 85% RTP [rated thermal
power]. However, the inspector identified that the 10 CFR 50.59 screen for
70 percent evaluation of this technical specification incorrectly stated that
SR [Surveillance Requirement] 3.3.1.11 is not required to be performed until
12 hrs after THERMAL POWER has reached or exceeded 85% RP [reactor
power].
(b) Technical Specification Surveillance Requirement 3.3.1.2 states, Verify total
Reactor Coolant System (RCS) flow rate as indicated by each CPC is less
than or equal to the RCS total flow rate. If necessary, adjust the CPC
addressable constant flow coefficients such that each CPC indicated flow is
less than or equal to the RCS flow rate, with a specified frequency of
12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />. Technical Specification Surveillance Requirement 3.3.1.5 states,
Verify total RCS flow rate indicated by each CPC is less than or equal to the
RCS flow determined by calorimetric calculations, with a specified
- 18 -
frequency of 31 days. Technical Specification Surveillance Requirement
3.3.1.2 and Surveillance Requirement 3.3.1.5 each contain a note stating
that the Surveillance Requirement is Not required to be performed until
12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> after THERMAL POWER > 85 percent RTP. However, the 50.59
screen for 70 percent evaluation of the technical specification did not
evaluate these surveillance requirements.
Nuclear Notification NN 202243314 was written to address these deficiencies.
Corrective actions included revising the 10 CFR 50.59 screening to evaluate the
actual wording of the Technical Specification Surveillance Requirement 3.3.1.11
frequency (Once after each refueling prior to exceeding 85% RTP,) which allows
performing the surveillance at 70 percent power. The 10 CFR 50.59 screening
change also added evaluation of Technical Specification Surveillance
Requirements 3.3.1.2 and 3.3.1.5, which stated that, although the surveillances
are not required to be performed until 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> after reaching 85 percent RTP, the
associated surveillance procedures had been revised to direct these surveillance
requirements be performed above 68 percent RTP. The surveillance procedures
were changed so the procedure steps do not conflict with the technical
specification wording. The licensee determined that changes to Technical
Specification Surveillance Requirements 3.3.1.2, 3.3.1.5 and 3.3.1.11 were not
required.
The inspectors noted that the original 10 CFR 50.59 written evaluation for this
change did not adequately evaluate the effect of limiting reactor power operations
to 70 percent on Technical Specification Surveillance Requirements 3.3.1.2,
3.3.1.5, and 3.3.1.11. Therefore the inspectors determined that the evaluation
was not adequate, in that it did not provide an adequate basis for the
determination that the change to limit reactor power operations to 70 percent did
not require a license amendment prior to implementing the change. Title10 CFR
50.59(d)(1) requires that the licensee maintain records of changes in the facility
that include a written evaluation which provides the bases for the determination
that the change, test, or experiment does not require a license amendment . . .
Contrary to the above, on September 14, 2012, the licensees written evaluation in
the 10 CFR 50.59 screening for ECP 800873488-0131 did not provide an
appropriate basis for the determination that the change to limit reactor power
operation to 70 percent did not require a license amendment.
Because this violation impacted the regulatory process, the inspectors assessed it
in accordance with the NRC Enforcement Policy, as directed by Inspection Manual
Chapter 0612, Appendix B, Issue Screening. The NRC Enforcement Manual
contains specific processes and guidance for implementing this Policy. NRC
Enforcement Manual, Part II, Section 2.1.3E.6.b states, in part, that minor
violations include the failure to meet 10 CFR 50.59 requirements that involve a
change to the final safety analysis description where there was no reasonable
likelihood that the change would ever require NRC approval per 10 CFR 50.59.
As described above, the change to limit reactor power operations to 70 percent
did not require changes to Technical Specification Surveillance Requirements
- 19 -
3.3.1.2, 3.3.1.5, and 3.3.1.11; thus, with respect to Part II, Section 2.1.3E.6.b of
the NRC Enforcement Manual, there is no reasonable likelihood that this would
ever require NRC approval. Therefore, in accordance with the NRC Enforcement
Manual, the inspectors determined that the failure to provide an adequate written
evaluation of the change to limit reactor power operation to 70 percent was a
minor violation of 10 CFR 50.59(d)(1).
(2)
Vibration and Loose Parts Monitoring System
(a) Updated Final Safety Analysis Report Section 5.4.1.5.5 states, Motor
vibration is sensed by the VLPM [vibration and loose parts monitoring] and
pump shaft vibration systems attached to the pump driver mount (motor
stand). Excessive vibration is alarmed in the control room.
(b) Updated Final Safety Analysis Report Table 7.6-4, Safety and Nonsafety
instrumentation Flooding Analysis, referenced the reactor coolant pump
P002/P004 vibration and loose parts monitoring transmitters.
The inspectors identified that Updated Final Safety Analysis Report
Section 5.4.1.5.5 and Table 7.6-4 were affected by the vibration and loose parts
monitoring system modifications and not addressed as part of the 50.59 screening
for Nuclear Engineering Change Package 800457837-0540. The licensee
evaluated the replacement vibration and loose parts monitoring system against
Regulatory Guide 1.133, Loose-part Detection Program for the Primary System of
Light Water Cooled Reactors, Revision 1, which specifies the monitoring
locations for the vibration and loose parts monitoring system detectors and
determined that the replacement of the Combustion Engineering system with the
Westinghouse Digital Metal Impact Monitoring System met the guidance specified
in Regulatory Guide 1.133. The licensee monitors reactor coolant pump vibration
using the reactor coolant pump shaft vibration system, which provides an alarm
indication in the control room, and the replacement of the vibration and loose
parts monitoring system does not affect the operation of the system. In addition,
there were no safety impacts associated with not updating Table 7.6-4, since the
changes involved abandoning equipment in place.
The inspectors observed the initial testing setup of the new Unit 2 vibration and
loose parts monitoring system. The inspectors reviewed the test results of the
Westinghouse Digital Metal Impact Monitoring System (vibration and loose parts
monitoring system) that included newly designed externally mounted sensors
located at the 7th tube support plate level of the replacement steam generators.
The test consisted of small impacts with a specially designed transducer hammer
inside the steam generator near the 7th tube support plate. The test results
demonstrated that the new system could detect the hammer impacts near the 7th
tube support plate level. However, the inspectors noted that the test results could
not be calibrated to detect tube-to-tube impact, since the test was limited to only
impacts to the tube support plate. There was not a test that could simulate tube
- 20 -
contact. Even with this limitation, the new external sensors were sensitive enough
to detect internal impacts at the upper level of the steam generator, specifically
associated with the 7th tube support plate.
Title 10 CFR 50.59(d)(1) requires that the licensee maintain records of changes in
the facility that include a written evaluation which provides the bases for the
determination that the change, test, or experiment does not require a license
amendment . . . Contrary to the above, on December 20, 2012, the licensees
Nuclear Engineering Change Package 800457837-0540 did not provide an
adequate basis for the determination that the change to the vibration and loose
parts monitoring system did not require a license amendment. Specifically, the
10 CFR 50.59 screening did not evaluate the effect of the vibration and loose
parts monitoring system modifications on Updated Final Safety Analysis Report
Section 5.4.1.5.5 and Table 7.6-4.
Because this violation impacted the regulatory process, the inspectors assessed it
in accordance with the NRC Enforcement Policy, as directed by Inspection Manual
Chapter 0612, Appendix B, Issue Screening. The NRC Enforcement Manual
contains specific processes and guidance for implementing this Policy. The NRC
Enforcement Manual, Part II, Section 2.1.3E.6.b, states, in part, that minor
violations include the failure to meet 10 CFR 50.59 requirements that involve a
change to the final safety analysis report where there was no reasonable
likelihood that the change would ever require NRC approval per 10 CFR 50.59.
As described above, the change to the vibration and loose parts monitoring
system modifications did not require a license amendment prior to implementing
the change so, with respect to Part II, Section 2.1.3E.6.b of the NRC Enforcement
Manual, there is no reasonable likelihood that this change would ever require
NRC approval. Therefore, in accordance with the NRC Enforcement Manual, the
inspectors determined that the failure to provide an adequate written evaluation of
the vibration and loose parts monitoring system modifications was a minor
violation of 10 CFR 50.59(d)(1). This deficiency was entered into the licensees
corrective action program as Nuclear Notification NN 202258050.
.5
(Closed) Unresolved Item 05000362/2012007-08, Non-Conservative Thermal-Hydraulic
Model Results
NRC Inspection Report 05000361/2012007 and 05000362/2012007described this
unresolved item, in part, as follows:
The team identified an unresolved item associated with the adequacy of
Mitsubishis FIT-III thermal-hydraulic code. The FIT-III code predicted
nonconservative low velocity and low void fraction results which were
used as inputs to the vibration code FIVATS. These non-conservative
thermal-hydraulic results led Mitsubishi to conclude that margins to
instability were significantly larger than they actually were.
- 21 -
During the original design, Mitsubishi used a number of computer codes to determine that
the design of the steam generators was adequate in regard to potential for vibration and
excessive wear. Based on key boundary operating parameters, primary temperature,
and flow rates from the SCE replacement steam generator design specification, the
Mitsubishi-developed steam generator steady-state performance code was used to
provide a one-dimensional thermal-hydraulic calculation of the overall steam generator
performance parameters, such as primary outlet temperature, steam pressure, steam
flow, and tube bundle circulation ratio.
The steady-state performance calculation code results were then used in the Mitsubishi
FIT-III thermal-hydraulic code to define velocity, density, and void fraction for input into the
vibration calculation. The Mitsubishi-developed FIVATS code was then used by
Mitsubishi in fluid-elastic stability analyses to determine if tubes were subjected to
conditions that would exceed their stability ratio design limit of 1.0, assuming that 1 of the
12 anti-vibration bar support contacts was ineffective.
It was determined during the Augmented Inspection Team inspections in March 2012 that
the SCE steam generators were under-designed in regard to margin to vibration and that
the lack of margin was largely due to under-prediction of gap velocity and void fraction by
the Mitsubishi FIT-III code analysis (Mitsubishi Document L5-04GA521, Three-
Dimensional Thermal and Hydraulic Analysis, Revision 3). The local thermal-hydraulic
analysis had a significant effect on the stability ratio results. It was concluded that the
U-bend velocities were underpredicted by a factor of 2.5 to 3. Additionally, the FIT-III
peak void fraction and quality were 0.95 and 0.67, respectively. The comparable values
from the ATHOS results were 0.996 and 0.91 at 100 percent power, respectively, where
fluid density is also 2.5 to 3 times smaller than the values computed by FIT-III. These two
under-predicted factors produced stability ratios that were lower by 20 to 40 percent as
compared to stability ratios based on ATHOS results. This resulted in no margin for a
small number of tubes and for the majority of tubes much less margin to the onset of
fluid-elastic instability than the designers or the licensee intended. Using Mitsubishi
design criteria of one inactive anti-vibration bar, there were some tube stability ratios that
exceeded 1.0.
a. Inspection Scope
The inspectors reviewed corrective action program documents and supporting
engineering evaluations associated with this unresolved item to determine if a
performance deficiency existed or if the issue constituted a violation of NRC
requirements. The inspectors reviewed Nuclear Notification NN 201836127 and the
associated cause evaluation performed by SCE to address the mechanistic cause of
the nonconservative results of the FIT-III thermal-hydraulic model and flow-induced
vibration analysis developed by Mitsubishi for the design of Units 2 and 3 replacement
steam generators. The inspectors also reviewed the status of SCEs and Mitsubishis
cause evaluation to identify the organizational and programmatic factors leading to
the nonconservative thermal-hydraulic model.
- 22 -
Additionally, the inspectors reviewed the replacement steam generator design
specification for the replacement steam generators to identify the applicable design
standards for thermal-hydraulic modeling and flow-induced vibration. The review of
design information included design basis documents for the original steam generators
to identify any design requirements for thermal-hydraulic modeling and flow-induced
vibration, in order to determine if those requirements were properly translated into the
replacement steam generator design specification. The inspectors reviewed the
applicable design standards to identify design information that would have prompted
the licensee to identify deficiencies in the thermal-hydraulic model and flow-induced
vibration analysis. Particularly, the inspectors reviewed the technical justification for
critical assumptions and design inputs.
The inspectors interviewed licensee staff and reviewed applicable quality assurance
requirements and site procedures for the verification of supplier documents to assess
whether the licensee had a reasonable opportunity to identify any deficiencies with
the thermal-hydraulic modeling and the flow induced-vibration analysis based on the
requirements and guidance in site procedures.
b. Findings
Introduction. For Unit 2, the NRC identified a Green noncited violation of 10 CFR
Part 50, Appendix B, Criterion III, Design Control, for the failure to verify the
adequacy of the thermal-hydraulic and flow-induced vibration design of the
replacement steam generators.
For Unit 3, the NRC identified an apparent violation of 10 CFR Part 50, Appendix B,
Criterion III, Design Control, for the failure to verify the adequacy of the thermal-
hydraulic and flow-induced vibration design of the replacement steam generators,
which also resulted in an associated apparent violation of Technical Specification
5.5.2.11, Steam Generator Program, because of a loss of tube integrity on Unit 3
Description. The effective construction code for SCE replacement steam generators
was the 1998 Edition, with 2000 Addenda, of the ASME BPVC,Section III. Article
NCA-3200 required that the owner shall prepare, review, and approve the
replacement steam generator design specification, which contains the specific design
requirements for the applicable code component. The replacement steam generator
design specification for SCE replacement steam generators (Document SO23-617-
01, Specification for Design and Fabrication of replacement steam generators for
Unit 2 and Unit 3, Revision 4), Section 3.8.2, stated that: The Supplier [MHI] shall
prepare and submit for SCEs approval a Performance Analysis Report documenting
all thermal-hydraulic aspects of the replacement steam generators. The Report shall
include all computer codes and modeling for the thermal-hydraulic performance of the
replacement steam generators. The Report shall include detailed calculations, by
region, showing that cross-flow velocities within the tube bundle shall be such as to
- 23 -
minimize tube wear at the tube to tube-support interfaces. The calculations shall
clearly identify the damping factor(s) used and margins to flow instability for steam
flow rates of up to 120% of the design flow rate.
Additionally, Section 3.21.7 of the replacement steam generator design specification
stated that, The Supplier shall provide a new thermal-hydraulic analytical model, or
update the existing plant EPRI ATHOS model for the replacement steam generators,
and furnish all input parameters required to update the existing steam generator
simulator model. The Supplier shall provide an executable version of the thermal-
hydraulic computer codes used in the design of the replacement steam generators.
Furthermore, Section 3.5.1 of the replacement steam generator design specification
stated that, To the extent practical, the version and identity of all Codes, Standards,
and other documents applicable to this Specification are shown in this Section.
Following this statement, the replacement steam generator design specification listed
ASME BPVC,Section III, Subsection NCA, and Division 1 Appendices as part of the
applicable standards.
Mitsubishi Document L5-04GA504, Evaluation of Tube Vibration, Revision 3,
adopted the methodology in Non-Mandatory Appendix N to ASME BPVC,Section III,
Dynamic Analysis Methods, to evaluate flow-induced vibration in tube arrays
exposed to cross flow. Specifically, Section 5 of this document stated that the
vibration analysis was performed in accordance with the procedures and suggested
inputs given in Appendix N-1330 to ASME Code Section III. This analysis was
reviewed and approved by SCE on January 28, 2008, during the design stage of the
replacement steam generators.
Paragraph N-1330 in Appendix N, to ASME Code Section III, provided
recommendations and inputs for avoiding fluid-elastic instability of tube arrays.
Paragraph N-1331.1, Prediction of the Critical Velocity, stated that the onset of
instability is governed, in part, by the flow velocity in the gaps between the tubes,
which is determined by Vg = Va x P/(P-D), where Vg is the gap velocity, Va is the
approach flow velocity that would occur if the tubes were not present, P is the tube
array pitch as defined in Figure N-1331-3, and D is the outside diameter of a tube.
In response to this unresolved item, Mitsubishi identified that one of the factors
responsible for the nonconservative flow velocities was that the flow area definition
was not consistent with the recommendations in Appendix N (Mitsubishi Document
L5-04GA591, Validity of Use of the FIT-III Results during Design, Revision 1).
Mitsubishi determined that the tube-to-tube gap used in the FIT-III thermal-hydraulic
code, to determine the gap velocities, was larger than the recommended value in
Appendix N, which resulted in lower calculated flow velocities. The difference in flow
area definition is illustrated below. The tube array in the SCE replacement steam
generators is a triangular array rotated 60 degrees, with a tube pitch of P = 1.0-inch.
For that type of array, ASME BPVC,Section III, Appendix N, Figure N-1331-3, defines
the tube pitch as the center-to-center distance between two tubes along the same
column/row and in the longitudinal direction of the flow, which in this case would be
- 24 -
P = 1.0 inch. However, the flow area defined in FIT-III used the tube pitch in the
transverse direction of the flow, which in SCE-rotated triangular array would be P =
1.73 inches. The use of a larger pitch in the FIT-III thermal-hydraulic analysis resulted
in nonconservative calculated (lower) flow velocities.
Mitsubishi Document L5-04GA521, Three-Dimensional Thermal and Hydraulic
Analysis, Revision 3, performed by Mitsubishi and approved by SCE during the
design phase of the replacement steam generators, showed that the thermal-
hydraulic model was built with two different pitch values. The report stated that the
model was built with a 1.0-inch pitch in the longitudinal direction of flow and 1.73-inch
pitch in the transverse direction. This analysis report was approved by SCE on
April 2, 2008. Additionally, the Evaluation of Tube Vibration report by Mitsubishi
stated that the thermal-hydraulic conditions for the FIT-III modeling were based on a
1.0-inch pitch in the longitudinal direction of flow and 1.73-inch pitch in the transverse
direction. These two design calculations were supporting documents for the
Performance Analysis Report required in the replacement steam generator design
specification. As indicated above, FIT-IIIs output for gap velocity results used the
1.73-inch distance instead of the 1.0-inch distance. The FIT-III code was developed
by Takasago, MHIs research and development center. Takasogo was responsible for
conducting the thermal-hydraulic analysis, using FIT-III, for each steam generator
design. The FIT-III results were then provided to the MHI Steam Generator Design
Flow
Flow
D = 0.75
PFIT-III =
1.73-inch
PASME = 1.0-inch
Vg = Va x P/(P-D)
- 25 -
Department, which input the gap velocity information into the FIVATS vibration code.
However, it was not recognized that the gap velocities input into the vibration code
were incorrect.
The inspectors determined that the licensee did not ensure that the thermal-hydraulic
modeling and flow-induced vibration analysis of the replacement steam generators
were adequate with respect to the replacement steam generator design specification.
Specifically, the licensee failed to ensure that the design calculations appropriately
incorporated the methodology from the ASME BPVC,Section III, Appendix N,
standard that was adopted by Mitsubishi for the flow-induced vibration analysis.
There were opportunities to identify this error during the early design stage of the
replacement steam generators. Licensee personnel questioned the analysis results
of FIT-III during design review meetings, but ultimately accepted the model results
and resultant design. From shortly after the contract was awarded until 2006, there
were letters, e-mails, meeting minutes, action item lists, and internal memoranda that
suggested concerns with all three of the elements that cause fluid-elastic instability,
which is void fraction, gap velocity, and adequacy of anti-vibration bar tube supports.
Regarding concerns raised about FIT-III gap velocities, Mitsubishi compared the
velocities to other Mitsubishi designed triangular pitch steam generators that also
used FIT-III, but did not compare the results to other similar-sized steam generators.
As a result of the failure to verify the adequacy of the thermal-hydraulic and flow-
induced vibration design, both Unit 3 replacement steam generators experienced
fluid-elastic instability in a localized area of the tube bundle leading to rapid,
significant, unexpected tube-to-tube wear. The tube degradation progressed to the
point of causing a primary-to-secondary leak in Steam Generator 3E0-88 through
Tube R106C78. Additionally, from March 13-21, 2012, the licensee conducted in-situ
pressure testing of the suspect tubes in both Unit 3 steam generators and identified a
total of eight tubes (including the leaking tube) that failed to meet the performance
criteria in plant Technical Specifications. The in-situ pressure testing identified that
Tubes R106C78, R102C78, R104C78, R100C80, R107C77, R101C81, R98C80, and
R99C81 in Steam Generator 3E0-88 failed to meet the structural integrity criterion in
Technical Specification 5.5.2.11. In addition to failing the structural integrity criterion,
Tubes R106C78, R102C78, and R104C78 also failed to meet the accident-induced
leakage criterion in Technical Specification 5.5.2.11.
Southern California Edison completed a review of the tube failures, including
conducting a deterministic root cause, an organization and programmatic root cause
(still ongoing), three different operational assessments, modification testing, and
submittal of a response dated October 3, 2012 (ML12285A263) to the NRCs
March 27, 2012, Confirmatory Action Letter (ML 12087A323). The organizational and
programmatic root cause evaluation has not been completed as of the issuance of
this report, in order to identify the causes of the breakdown in design control such that
comprehensive corrective actions can be taken to not only prevent recurrence, but
prevent the failures of other important structures, systems, and components that may
be subject to the same or similar design problems.
- 26 -
Unit 2:
Analysis. The inspectors determined that the licensees failure to verify the adequacy
of the thermal-hydraulic and flow-induced vibration design of the replacement steam
generators was a performance deficiency. Criterion III specifies that design control
measures shall provide for verifying or checking the adequacy of design, in particular,
thermal and hydraulic analyses. This performance deficiency is more than minor, and
therefore a finding, because it is associated with the equipment performance attribute
of the Initiating Event Cornerstone and adversely affected the cornerstone objective of
limiting the likelihood of those events that upset plant stability and challenge critical
safety functions during shutdown as well as power operations.
The inspectors used NRC Inspection Manual Chapter 0609, Attachment 4 and
Appendix A, to evaluate the significance of this finding. In accordance with Exhibit 1
of Inspection Manual Chapter 0609, Appendix A, the inspectors determined that the
finding is of very low safety significance because the finding did not involve a
degraded steam generator tube that could not sustain three times the normal
operating differential pressure and did not violate the accident leakage performance
criterion.
The licensee initiated Nuclear Notification NN 202447268 to address this issue in the
corrective action program and implement corrective actions to prevent recurrence.
Southern California Edison revised the thermal-hydraulic code of record and ensured
that the code was in accordance with ASME guidance.
No crosscutting aspect was assigned because this performance deficiency occurred
in the 2005 to 2008 timeframe. Substantial management and personnel changes
have occurred, including taking actions to address a chilled work environment and
various crosscutting themes. The NRC determined that the performance behavior
that existed at that time is not indicative of current performance.
Enforcement. Title 10 CFR Part 50, Appendix B, Criterion III, Design Control,
requires, in part, that design control measures shall be established to provide for
verifying or checking the adequacy of design, such as by the performance of design
reviews, by the use of alternate or simplified calculational methods or by the
performance of a suitable testing program.
Contrary to the above, on January 28, 2008, and April 2, 2008, SCE failed to verify or
check the adequacy of Mitsubishis developed design Documents L5-04GA504
(SO23-617-1-C157), Evaluation of Tube Vibration, Revision 3, and L5-04GA521
(SO23-617-1-C683), Three-Dimensional Thermal and Hydraulic Analysis,
Revision 3, respectively, for the flow-induced vibration and thermal-hydraulic designs.
Specifically, the output of the thermal-hydraulic code and input to the vibration code
were not verified or checked to be in accordance with ASME Section III, Appendix N,
Dynamic Analysis Methods. Because the finding is of very low safety significance
and has been entered into the licensees corrective action program as Nuclear
Notification NN 202447268, this violation is being treated as a noncited violation
- 27 -
consistent with Section 2.3.2 of the NRC Enforcement Policy: NCV
05000361/2012009-01, Failure to Verify Adequacy of Thermal-Hydraulic and Flow-
Induced Vibration Design for the Unit 2 Replacement Steam Generators.
Unit 3:
Analysis. Regarding Unit 3, this failure also constitutes a performance deficiency for
the same reason previously discussed for Unit 2. Specifically, the failure to verify the
adequacy of the thermal-hydraulic and flow-induced vibration design resulted in
significant and unexpected steam generator tube wear due to fluid-elastic instability,
which challenged the structural integrity of the steam generator tubes to perform their
pressure boundary function.
The inspectors used NRC Inspection Manual Chapter 0609, Attachment 4 and
Appendix A, to evaluate the significance of this finding. In accordance with Exhibit 1
of Inspection Manual Chapter 0609, Appendix A, the inspectors determined that this
finding required evaluation in accordance with Inspection Manual Chapter 0609,
Appendix J, because the finding involved a degraded steam generator tube condition,
where one tube cannot sustain three times the differential pressure across a tube
during normal full power, steady-state operation. In accordance with Inspection
Manual Chapter 0609, Appendix J, this finding required a detailed risk analysis, since
it involved two or more tubes that could not sustain three times the normal differential
pressure and one or more steam generators that violated accident-induced leakage
performance criterion. A Phase 3 analysis was completed using the San Onofre
SPAR model, Revision 8.22, assuming average test and maintenance, and a
truncation limit of 1.0E-11. Based on the best available information, the performance
deficiency was preliminarily characterized as a finding of low to moderate safety
significance (White). Refer to Attachment 4 for the detailed Phase 3 analysis.
The licensee initiated Nuclear Notification NN 202447265 to address this issue in the
corrective action program. Southern California Edison revised the thermal-hydraulic
code of record and ensured that the code was in accordance with ASME guidance.
No crosscutting aspect was assigned because this performance deficiency occurred
in the 2005 to 2008 timeframe. Substantial management and personnel changes
have occurred, including taking actions to address a chilled work environment and
other safety culture issues. The NRC determined that the performance behavior that
existed at that time is not indicative of current performance.
Enforcement. Title 10 CFR Part 50, Appendix B, Criterion III, Design Control, states,
in part, that design control measures shall be established to provide for verifying or
checking the adequacy of design, such as by the performance of design reviews, by
the use of alternate or simplified calculational methods or by the performance of a
suitable testing program.
Technical Specification 5.5.2.11, Steam Generator Program, Section b,
Performance criteria for SG tube integrity, states, in part, that steam generator tube
- 28 -
integrity shall be maintained by meeting the performance criteria for tube structural
integrity and accident induced leakage. Technical Specification 5.5.2.11 b.1,
Structural integrity performance criterion, states, in part, that this includes retaining a
safety factor of 3.0 against burst under normal steady-state full power primary-to-
secondary differential pressure. Technical Specification 5.5.2.11 b.2, Accident
induced leakage performance criterion, states, in part, that leakage shall not exceed
0.5 gallons per minute per steam generator for a main steam line break accident.
Contrary to the above, on January 28 and April 2, 2008, SCE failed to verify or check
the adequacy of Mitsubishis developed design Documents L5-04GA504 (SO23-617-
1-C157), Evaluation of Tube Vibration, Revision 3, and L5-04GA521 (SO23-617-1-
C683), Three-Dimensional Thermal and Hydraulic Analysis, Revision 3, respectively,
for the flow-induced vibration and thermal-hydraulic designs. Specifically, the output
of the thermal-hydraulic code and input to the vibration code were not verified or
checked to be in accordance with ASME Section III, Appendix N, Dynamic Analysis
Methods.
Consequently, the inadequate thermal-hydraulic and flow-induced vibration design
resulted in adverse flow conditions, along with insufficient tube support, which caused
fluid-elastic instability of a group of tubes in both Unit 3 replacement steam
generators. This resulted in one tube leaking and required operator response to
rapidly shut down Unit 3 on January 31, 2012. In March 2012, in-situ pressure testing
on Unit 3 Steam Generator 3E0-88 confirmed that eight steam generator tubes failed
to meet the performance criterion for structural integrity and three of those tubes also
failed to meet the accident-induced leakage criterion. During in-situ pressure testing,
Tubes R106C78, R102C78, R104C78, R100C80, R107C77, R101C81, R98C80, and
R99C81 in Steam Generator 3E0-88 failed to meet the structural integrity criterion
limit of three times the normal steady-state primary-to-secondary differential pressure
of 5250 psig (room temperature equivalent to 4290 psi under hot 100 percent power
conditions), with the tubes failing at test pressures ranging from 2874 psig to
5026 psig (at room temperature). In addition, Tubes R106C78, R102C78, and
R104C78 failed to meet the accident-induced leakage criterion of not exceeding
0.5 gpm leakage per steam generator at a main steam line break test pressure of
3200 psig (room temperature equivalent to 2560 psig differential pressure during main
steam line break), with each tube having leakage rates of approximately 4.5 gpm,
prior to exceeding 3200 psig. Because this finding has been preliminarily determined
to be of low-to-moderate safety significance (White), it will be treated as an apparent
violation and tracked as AV 05000362/2012009-02, Failure to Verify Adequacy of
Thermal-Hydraulic and Flow-Induced Vibration Design for the Unit 3 Replacement
- 29 -
.6
(Closed) 05000362/2012007-04: Evaluation of Changes in Dimensional Controls during
the Fabrication of Unit 2 and Unit 3 Replacement Steam Generators
a. Inspection Scope
NRC Inspection Report 05000361/2012007 and 05000362/2012007 described this
unresolved item, in part, as follows:
Based on the information gathered by the team on the differences in
dimensional controls of critical parameters in Unit 2 and Unit 3
replacement steam generators, the team determined that Mitsubishi
did not consider the potential impact of improving dimensional
controls for tube roundness and anti-vibration bars on the final tube
bundle clearances at normal operating conditions.
The inspectors reviewed the following reports, which assessed the differences in
dimensional controls between SCE Units 2 and 3, and the impact of these differences
on the tube-to-anti-vibration bar gap distributions and tube-to-anti-vibration bar
contact force distributions throughout the U-bend region and their effect on fluid-
elastic instability performance at Units 2 and 3.
Mitsubishi Document L5-04GA564, Tube Wear of Unit-3 RSG - Technical
Evaluation Report (Attachment 4 to the San Onofre Nuclear Generating Station
Unit 2 Return to Service Report)
AREVA Document1814-AU651-MO160, SCE Unit 2 Cycle 17 Steam Generator
Operational Assessment for Tube-to-Tube Wear, Revision 0
The Mitsubishi/AREVA assessment of the dimensional control differences was
performed as part of the SCE operational assessment. The inspectors also met with
cognizant Mitsubishi and AREVA personnel to discuss this assessment.
b. Observations and Findings
No findings were identified.
The inspectors determined that Mitsubishi relied on industry standards and guidance
during the fabrication and design of the replacement steam generators. Dimensional
controls for tube roundness and anti-vibration bars were identified by Mitsubishi as
conservative on the final tube bundle clearances at normal operating conditions.
The inspectors assessed whether the reported changes to the dimensional controls of
the upper bundle structure were properly evaluated during the fabrication of the
Units 2 and 3 replacement steam generators. These changes were within the design
specifications and included manufacturing changes to the anti-vibration bars and
steam generator tube, which were meant to reduce the wear between anti-vibration
bar and tubes. As documented in NRC Augmented Inspection Team Report
05000361/2012007;362/2012007, the replacement steam generators were required to
- 30 -
be designed, fabricated, and tested in accordance with the "Conformed Specification
for Design and Fabrication of the Replacement Steam Generators," also known as the
design specification, and contained identical technical requirements for Units 2 and 3
steam generators. All replacement steam generators were required to be designed,
fabricated, and tested in accordance with the 1998 edition of the American Society of
Mechanical Engineers (ASME) Boiler and Pressure Vessel Code, with the 2000
Addenda, industry standards, and NRC endorsed methods described in applicable
regulatory guides. The inspectors determined that the design specifications
contained the same requirements for both the Units 2 and 3 replacement steam
generators.
Mitsubishis Document L5-04GA564, Tube wear of Unit-3 RSG - Technical
Evaluation Report, Revision 5, determined that improvement of dimensional controls
associated with anti-vibration bar flatness was the key contributing cause in
determination of the failure mechanism leading to tube-to-tube wear in Unit 3. The
inspectors noted that the specifications also referenced inputs from industry guidance
found in Tubular Exchanger Manufacturers Association (TEMA), 8th Edition, 1999.
The inspectors noted that the TEMA guidance permitted practical use of engineering
principles and field experience in the manufacturing and design of the tube heat
exchangers. The general scope of the TEMA standard clearly indicates that
damaging tube vibration can occur under certain conditions, including unsupported
tube spans. The replacement steam generator design specification input used the
TEMA guidance in determining the maximum unsupported tube spans. In cases
where engineering analyses showed the probability of destructive vibration, the
designer was directed to analyze key design conditions, including the thermal-
hydraulic limitations, baffle design, and tube span. This was necessary to account for
and prevent flow-induced vibration damage.
Mitsubishi organizational and programmatic root cause analysis Report UES-
20120254, Root Cause Analysis Report for Tube Wear Identified In The Unit 2 and
Unit 3 Steam Generators of San Onofre Nuclear Generating Station, Revision 0,
concluded that the replacement steam generators thermal-hydraulic conditions (void
fractions) were high but designers performed feasibility studies to improve them. The
report stated that several design adjustments were considered to reduce the adverse
thermal-dynamic conditions, but the effects were small. Therefore, the designers
concluded that the final design was adequate. The inspectors noted that the
Mitsubishi cause evaluation stated that in-plane fluid-elastic instability was a
phenomenon that had not been experienced in nuclear U-tube steam generators and
that the upper bundle tube-to-anti-vibration bar gaps were within design
specifications. However, the root cause identified that tube-to-tube wear was caused
by large displacements of tubes in the in-plane direction because of ineffective anti-
vibration bar supports.
As identified by the cause evaluation, prior to the design phase of SCE replacement
steam generators, the accepted industry U-bend design practice applied flat bar (anti-
vibration bar) supports because of significant advantages such as decreased tube-to-
anti-vibration bar wear rates. Prior to 2005, it had been an established practice to
- 31 -
consider anti-vibration bar or flat bar supports very effective in U-Bend steam
generators. However, the inspectors noted that earlier industry studies had
documented that flow-induced vibration characteristics were not well understood. In
fact, in a 1983 study referenced in the SCE Westinghouse operational assessment
titled, The Effect of Flat Bar Supports on the Cross flow Induced Response of Heat
Exchanger U-Tubes, in the Journal of Engineering for Power, October 1983,
page 27, it stated the need for additional research, since supports (flat bars) may not
adequately restrict in-plane tube motions.
The NRC reviewed the Mitsubishi technical evaluation report on Unit 3 replacement
steam generator tube wear and noted that dimensional controls were reportedly
improved. This improvement was in standard deviation of Unit 3 anti-vibration bar
dimensions. The inspectors evaluated the information, which shows a small
improvement in standard deviations associated with only the bending portion of anti-
vibration bar thickness change from nominal thickness. The straight bar section
standard deviation was not notably better between Units 2 and 3 replacement steam
generators. It was noted that differences in anti-vibration bar twist between Units 2
and 3 were determined to be the biggest contributor to differences in the tube
bundles. The inspectors also reviewed the results of another comparative analysis
based on visual inspection and review of Document L5-04GA564, Appendix 7, Visual
Inspection Results for U-Bend Region for Unit-2/3. The licensee concluded that,
even though Unit 3 operated for a much shorter time compared to Unit 2, there were
no obvious visual differences, but there were differences in the wear patterns. For
example, the visual observations from Units 2 and 3 did not show large gaps between
the anti-vibration bars and tube. The anti-vibration bars appeared to be straight, with
no detectable abnormalities with weld caps or upper structure orientation. Wear
patterns in Unit 2, which had operated the longest, did not have the wear pattern seen
in Unit 3 steam generators, which showed evidence of extended tube wear scaring
attributed to in-plane motion or vibration, with evidence of orbital tube movement
relative to the anti-vibration bar and tube.
The inspectors concluded that Mitsubishi followed its design specifications as
required.
.7
Resolution of Independent Technical Review Findings
a. Inspection Scope
The inspectors reviewed the findings prepared by NRC consultants in a report,
Independent Evaluation of San Onofre Nuclear Generating Station (SCE) Steam
Generator Tube Wear Problems, dated July 13, 2012 (see Attachment 2 of this
report). The consultants findings were based on their review of documents available
at that time, including a draft copy of NRC Augmented Inspection Team Report
05000361/2012007 and 05000362/2012007 that was issued on July 18, 2012.
b. Observations and Findings
No findings were identified.
- 32 -
The NRC consultants made a number of observations during their independent
review of the scope and effort of the Augmented Inspection Team. This inspection
addressed each observation as follows:
Consultant observation: Review and assess the inconsistencies between
replacement steam generator Design L5-04GA510, Thermal and Hydraulic
Parametric Calculations, Revision 5, and the revised calculated thermal-hydraulic
performance.
NRC inspection result: The inspectors determined that the steady-state
performance code was built based on the design inputs and other requirements
provided by the design specification and on the geometrical characteristics and
internal components of the Mitsubishi replacement steam generators. The
steady-state performance code was run first and it provided all the basic operating
parameters and performance information. The analysis provided the main
operating parameters, including saturation pressure, circulation ratio, steam
flowrate, tube side pressure drop, feedwater pressure at the feedwater inlet
nozzle, all individual component pressure losses in the circulation loop, global
heat transfer coefficient, and the water and steam inventories. The calculation
results are shown in Document L5-04GA510, Table 6.1.5-1. The data at
beginning-of-life operating conditions for a Thot of 598oF were used as Cycle 16
operating conditions that were input into the original FIT-III analysis and recent
ATHOS base case analysis. NRC staff noted that all the ATHOS models, i.e., the
Mitsubishi model, the Westinghouse model, and the NRC independent model,
used these steady-state performance code results as input.
Since the purpose of the steady-state performance code was only to determine
the global parameters, it was not necessary that local phenomena, such as
subcooled boiling height in the bundle, match with the ATHOS analysis. The
performance code results were used as boundary conditions so that the ATHOS
analysis could define the local node-by-node thermal-hydraulic conditions that
were needed for the vibration analysis. Additionally, in the ATHOS analysis, the
subcooled boiling height varied by tube row based on tube temperature and local
flow conditions. In conclusion, there were no major issues with this document or
the performance results provided.
The key results of the FIT-III analysis were provided in Document L5-04GA521,
Figure 8.1-2a, showing a predicted maximum steam quality of 0.67 and maximum
void fraction of 0.95. The conditions postulated considerably underestimated the
steam quality and, therefore, also underestimated steam velocities and void
fraction.
The outputs of the FIT-III analysis were used in the vibration analysis of essential
components, primarily the tubes. The vibration design was contained in
Document L5-04GA504, Evaluation of Tube Vibration, Revision 3. In this
analysis, the Mitsubishi FIVATS code was used to evaluate the design for tube
- 33 -
vibration to justify the number and layout of the tube support plates and anti-
vibration bars proposed. The Mitsubishi design methodology used the Connors
Equation to evaluate out-of-plane flow instability using typical ASME suggested
design values for K of 2.4 and h of 1.5 percent in the U-bend region.
Table 2-4 of Document L5-04GA504 showed the limiting stability ratio results,
assuming one anti-vibration bar support point was inactive, indicating that the
maximum expected stability ratio was 0.54. These stability ratio results were
considerably low with respect to the ASME acceptance criteria of less than 1.0.
Consultant observation: The observed difference in standard deviation values for
tube diameter on the U-bend flanks (sometimes referred to as G-value or
ovality in the consultants and Mitsubishis reports) as reported in Document L5-
04GA564, Tube wear of Unit-3 RSG - Technical Evaluation Report, Revision 2,
was considered by the consultants to have minimal effect on any difference in the
contact forces at the tube-to-anti-vibration bar supports between Units 2 and 3,
particularly if one takes into consideration that the low radius U-bends are the
biggest contributor to tube ovality.
NRC inspection result: The inspectors reviewed the tube diameter data in
Document L5-04GA564, Revisions 2 and 9, and concur with this observation for
the same reasons cited by the consultants. The inspectors noted that this
observation was consistent with the results of the Mitsubishi manufacturing
dimensional dispersion analysis in Document L5-04GA564, Revision 9,
Appendix 9, which shows that the mean differences in G-values from nominal and
differences in G-value standard deviations from the mean between Units 2 and 3
are small compared to the differences in anti-vibration bar twist between Units 2
and 3, which are the dominant contributor to the higher contact forces (tighter tube
bundle) being calculated by Mitsubishi for Unit 2 versus Unit 3.
Consultant observation: Absent the existence of additional information, there is
no apparent basis to believe that the number of local radius adjustments during
manufacture of the U-bends, as reported in Mitsubishi Document L5-04GA564,
Revision 2, has any relevance to the observed steam generator tube degradation.
NRC inspection result: The inspectors reviewed the U-bend radius adjustment
data in Mitsubishi Document L5-04GA564, Revisions 2 and 9, and concur with this
observation. Local radius adjustments are sometimes needed to bring the U-bend
profiles to within the required specifications. However, the inspectors found that
U-bend radius variability did not directly impact tube-to-anti-vibration bar gaps,
provided the profile requirements were met. In addition, the inspectors noted that
this observation was consistent with the results of the Mitsubishi manufacturing
dimensional dispersion analysis in Document L5-04GA564,
Appendix 9, Revision 9, which did not consider U-bend radius variability a relevant
parameter affecting the tube-to-anti-vibration bar gap and contact force
distributions.
- 34 -
Consultant observation: The average of the gaps between the outermost tubes
and the central columns was found to be essentially the same between the
Units 2 and 3 steam generators, which does not support a premise that more
uniform manufacturing practices for the Unit 3 tube bundles resulted in less
contact force between the tubes and anti-vibration bars. In the absence of more
dimensional information for the steam generator tube bundles, it is not believed
possible to explicitly define the number of active supports in the Units 2 and 3
NRC inspection result: The inspectors reviewed the gap information (Mitsubishi
Document L5-04GA564, Revision 2) cited by the consultants in addition to
Revision 9 of the same report. The inspectors concur that the measured gaps
between the outermost tubes and the anti-vibration bars in the central columns do
not in-and-of-themselves support a premise that more uniform manufacturing
practices for the Unit 3 tube bundles resulted in less contact force between the
tubes and anti-vibration bars. The inspectors noted, however, that conclusions by
SCE relating to the number of active supports in the Units 2 and 3 steam
generators are based on analyses documented in Mitsubishi Document L5-
04GA564, Revision 9, Appendix 9, and in SCE Document 1814-AU651-MO160,
SONGS Unit 2 Cycle 17 Steam Generator Operational Assessment for Tube-to-
Tube Wear, Revision 0, prepared by AREVA. These analyses were not reviewed
by the consultants, but were reviewed as part of the NRR technical evaluation and
by the inspectors; however, the NRC review was not completed because of the
decision by SCE to permanently cease operation of Units 2 and 3.
Consultant observation: Eddy current test inspection measurements of the tube-
to-anti-vibration bar gap were determined to be of questionable value in an
assessment of likely tube wear behavior. Review of Figure 4.1.2-1 in Mitsubishi
Document L5-04 GA564, Revision 2, indicated the potential fallacy in projecting
differences in average contact forces (at tube-to-anti-vibration bar intersections)
between Units 2 and 3. Specifically, Figure 4.1.2-1 shows virtually identical
average absolute signal amplitudes at anti-vibration bar locations for Steam
Generators 2E0-88 and 3E0-89 that have shown significant differences in
operational tube wear behavior.
NRC inspection result: The inspectors reviewed the Mitsubishi data in both
Revisions 2 and 9 of L5-04GA564 and concur that this data lends little insight as
to the significantly different amounts of tube-to-tube wear observed between
Units 2 and 3. The Mitsubishi signal amplitude data is based on bobbin probe
data. Since the time of the consultant review, AREVA conducted extensive gap
measurements using eddy current pancake and ultrasonic techniques expected
by AREVA to provide a more accurate and comprehensive gap assessment. The
purpose of these measurements was to determine if there were highly
heterogeneous spatial distributions of large gaps among the steam generators
that might affect the results of the AREVA operability assessment.
- 35 -
Consultant observation: Review the potential cause for low flow velocities. There
is a region of almost stagnant flow, possibly caused by higher flow resistance for
the cross-flow from the wrapper inlet ports into the tube bundle due to the smaller
pitch-to-diameter ratio of the replacement steam generators than in the original
steam generators. A comparison of the replacement steam generator thermal-
hydraulics with that of the original steam generator was not found in either the
Augmented Inspection Team or SCE root cause reports, which could aid in the
determination of the cause for the flow abnormalities in the replacement steam
generator.
NRC inspection result: The inspectors determined that flow in the downcomer
was restricted by protrusions of inspection ports, wrapper supports, and tube
support plate anti-rotation blocks. At the bottom of the downcomer (top of
tubesheet), flow is forced to make a 90-degree turn and is directed into the tube
bundle. This flow is affected by the height of the wrapper opening, size of the
open lane, and pitch-to-diameter ratio.
As the flow continues in the bundle, it slows considerably as it begins upflow in
the bundle. There are some places on top of the tubesheet that should be
evaluated for low flow and for potential sludge accumulation. It is not uncommon
to see areas of low flow, so modern replacement steam generators include
features for 1-2 percent blowdown flow and enhanced sludge lancing. There also
are slightly more tubes in the replacement steam generator, but not enough to
have any effect on sludge potential or any flow concerns in the first span.
There is no thermal-hydraulic model associated with the original steam
generators, since these were designed prior to the development of the ATHOS
model in 1985; therefore, the inspectors were not able to compare the flows.
Consultant observation: The correlation between boiling in a small region at the
bottom of the tube bundle, based on Mitsubishi document calculations, and the
observed region of tube wear increasing from tube support Plate 1 levels upward
into the U-bend region, has apparently not been addressed. Some mechanism is
moving the tubes and causing the tube-to-tube support plate wear in a small
region.
NRC inspection result: There is potential of subcooled nucleate boiling in lower
tube support plates in lower rows on the hot side of the steam generators. The
effects of this nucleation are usually not significant, so it is quite often not
considered during design of steam generators. Steam generators with lower
recirculation ratios, like the replacement steam generators, are more susceptible
because the tube bundle contains a larger percentage of steam than other steam
generator designs; however, the likelihood of this causing tube movement would
be considered inconsequential. It should be noted that the pitch-to-diameter ratio
in the vertical section of the bundle is the same (1.33) in the replacement steam
generator and original steam generator. However, in the U-bend area,
- 36 -
incrementation (indexing) was used after row 72, which resulted in an increased
pitch-to-diameter ratio of up to a maximum of approximately 1.53 for the
outermost tube row.
Consultant observation: Review the simplified scenario of Section 6.3 and make a
determination as to the validity. From the report, More evaluations would be
required to substantiate the postulated scenario as the source of the high void
fraction and velocity in a specific U-bend region.
NRC inspection result: The inspectors developed their own independent ATHOS
thermal-hydraulic model and reviewed three other thermal-hydraulic models, all
yielding relatively similar results. The codes used include the EPRI ATHOS code,
Westinghouse modified ATHOS code, and French CAFCA4 code. Each of the
codes used homogeneous methods with empirical models for heat and mass
transfer and drift flux models to compute two-phase flow conditions. The code
methods date to the 1970s and 1980s, but they have been widely used and
benchmarked to available scaled and full-plant data. Additionally, the ATHOS and
CAFCA4 code results have been successfully used to design many replacement
Modeling two-phase flows is very complex since it exhibits various flow regimes,
or flow patterns, depending on the void fraction of the two-phase fluid and the flow
rate. Additionally, flow patterns can be irregular or chaotic. However, averaged
behavior based on conservation equations can be used to model one-dimensional
steady-state and transient two-phase flow in reactors and steam generators.
Some simple transients are relatively easy to model and can be validated with
data while others are more difficult and data to validate the results are scarce.
Three-dimensional analysis with axial, radial, and tangential control volumes adds
additional complexities of momentum equations for the added directions. The
three-dimensional code uses a porous media approach to represent local
geometries and requires customized pre-processors to model modern designs of
anti-vibration bars and tube support plates.
Westinghouse maintains its own version of ATHOS and has completed the most
comprehensive validation of the code. To support plant restart, Mitsubishi was
directed by SCE to use the ATHOS, with their analyses being reviewed by AREVA
and other SCE consultants. The NRC reviewed the benchmarking and validation
of the ATHOS code to actual steam generator conditions as follows:
The Westinghouse validation suite includes:
(1)
Model Boiler Number 2 (MB-2) one percent power-scaled model of the
Westinghouse Model F steam generator, designed to be geometrically and
thermal-hydraulically similar to the Model F, and capable of generating
10 MWt of power. The model was able to produce dry saturated steam at
1000 psia, the same as with Model F.
- 37 -
(2)
Full-scale steam generator data measured at Electricity of France (EDF)
nuclear power plants Bugey 4 and Tricastin 1 (Westinghouse Model 51A
and 51M) operated at full and reduced power levels.
(3)
EPRI full-scale steam generator test data collected from operating
Westinghouse Model F and D4 steam generators.
(4)
French Alternative Energies and Atomic Energy Commission (CEA) Clotaire
scaled test program with the main focus on thermal-hydraulic data collected
on the secondary side fluid void fraction and axial vapor velocity throughout
the tube bundle. This data had not been previously verified. Nine
organizations from six countries participated in this program to verify six
different three-dimensional thermal-hydraulic codes developed for
pressurized steam generator analysis. Westinghouse participated to verify
the ATHOS code.
(5)
Leonard Cold Flow Test using a scaled down model of the U-bend region of
a bundle similar to the Westinghouse Model 51 design with three test
configurations: (a) no anti-vibration bars, (b) two sets of anti-vibration bars,
and (c) three sets of anti-vibration bars. Experiments were conducted in
each configuration to obtain velocity distributions along the U-bends.
From these test cases, Westinghouse concluded that the ATHOS code calculates
thermal-hydraulic parameters and behavior in good agreement with measured
data.
These experimental data and test parameter matrices appear to adequately
encompass the regions and conditions of concern for Cycle 16 operation of SCE
replacement steam generators. The code has wide use and acceptance in the
nuclear industry for analysis of recirculating-type steam generators. As with any
thermal-hydraulic code, the accuracy of the results is a function of the code and
the ability of the user to correctly model and interpret the output. The inspectors
believe that ATHOS code modeling is sufficiently capable of representing these
thermal-hydraulic conditions, and the users were sufficiently proficient in building
the input models.
The results of several models were reviewed to establish a level of confidence in
the key ATHOS output for vibration analysis for two tubes in the tube-to-tube wear
affected region of the bundle. Results reviewed include the NRC independent
ATHOS thermal-hydraulic analysis, the Mitsubishi ATHOS thermal-hydraulic
analysis, the Westinghouse in-house ATHOS60 thermal-hydraulic analysis, and
the AREVA CAFCA4 thermal-hydraulic analysis. Comparisons are shown for gap
velocity, void fraction, and fluid density along the U-bend of the subject tubes at
100 percent power operation. As indicated, the major differences in the results
are located on the hot side of the U-bend, where the tubes are hotter and there is
- 38 -
more bulk boiling. Additionally, the wrapper transition opening is located in this
region so there is an increase flow area and, consequently, increases in fluid
velocities along the entire periphery.
The NRC and Mitsubishi results show the higher peak void fractions, which also
peak at a lower location along the U-bend as compared to the Westinghouse and
AREVA results. At higher angles along the U-bend, the void fraction predicted by
each code tends to merge, and the remainder along the cold side tends to show
very similar trends. As expected, comparisons of fluid density show similar trends
as the void fraction. Dry saturated steam at about 850 psia has a density of about
2.0 lbm/ft3, and the NRC, Mitsubishi, and Westinghouse results predict that there
are significant areas in the U-bend, in the affected region, where velocities are
high and steam is nearly dry. Peak velocities and void fraction support the tube-
to-tube wear patterns found in the Unit 3 steam generators. The inspectors
concluded that the ATHOS code models are adequate to predict the approximate
location and magnitude of the high void fraction and velocity, which are the key
contributors to tube-to-tube wear found in the U-bend region.
Figure 4: U-Bend Gap Velocities
Tube R100C80
Tube R95C85
\\
0
5
10
15
20
25
30
35
0
20
40
60
80
100
120
140
160
180
Velocity (ft/sec)
U-Bend Angle (degrees)
NRC ATHOS
MHI ATHOS
WEC ATHOS
Areva CAFCA4 -
[R106C78]
0
5
10
15
20
25
30
35
0
20
40
60
80
100
120
140
160
180
Velocity (ft/sec)
U-Bend Angle (degrees)
NRC ATHOS
MHI ATHOS
WEC ATHOS
Areva CAFCA4 -
[R94C88]
- 39 -
Figure 5: U-Bend Void Fraction
Tube R100C80
Tube R95C85
F
Figure 6: U-Bend Density
Tube R100C80
Tube R95C85
0
5
10
15
20
25
30
0
20
40
60
80
100
120
140
160
180
Density (lbm/ft3)
U-Bend Angle (degrees)
NRC ATHOS
MHI ATHOS
WEC ATHOS
Areva CAFCA4 -
[R106C78]
0
5
10
15
20
25
30
0
20
40
60
80
100
120
140
160
180
Density (lbm/ft3)
U-Bend Angle (degrees)
NRC ATHOS
MHI ATHOS
WEC ATHOS
Areva CAFCA4 -
[R94C88]
0.7
0.75
0.8
0.85
0.9
0.95
1
0
20
40
60
80
100
120
140
160
180
Void Fraction
U-Bend Angle (degrees)
NRC ATHOS
MHI ATHOS
WEC ATHOS
Areva CAFCA4 -
[R106C78]
0.7
0.75
0.8
0.85
0.9
0.95
1
0
20
40
60
80
100
120
140
160
180
Void Fraction
U-Bend Angle (degrees)
NRC ATHOS
MHI ATHOS
WEC ATHOS
Areva CAFCA4 -
[R94C88]
- 40 -
.8
Chalk River Testing
The purpose of the Chalk River testing was to support the SCE Unit 2 return-to-service
steam generator operational assessment analyses for Unit 2 as related to the Unit 3 tube
leak and loss of tube integrity due to tube-to-tube wear in the upper bundle. One of the
analytical methods contained in a number of the operational assessments used an
empirical damping correlation, called squeeze film damping, based on testing done in
1988 for a tube and drilled support plate arrangement. During the NRCs review of the
operational assessments and their application of squeeze film damping, the inspectors
had a number of questions related to the applicability of the data since the 1988 testing
configuration was significantly different from the flat-bars (anti-vibration bars), and the test
range did not include lower frequencies where this information was being applied.
Atomic Energy Canada Limited performed a series of tests with a straight steam
generator tube under a variety of support and excitation conditions, including the type of
support, tube-to-support gaps, excitation levels, tube-to-support impacts, and tube-to-
support preload. The tests were performed in air and still water at four different tube
lengths corresponding to vibration frequencies in water between 4.5 Hz and 32 Hz. The
primary objective of the tests were to obtain damping data at low frequency (<20 Hz) with
flat bar (anti-vibration bar type) supports.
a. Inspection Scope
The inspectors reviewed design specifications of the test rig, test rig setup,
calculations, procedures, and comparison of original test rig data to revised test rig
data. In some instances, the inspectors performed independent calculations to verify
the testing results. The inspectors verified that the condition of the components was
consistent with the design; verified that the test measuring devices were appropriately
calibrated and of the correct range; reviewed equipment dedication; observed a
number of actual tests, with varying configurations; and independently verified that the
test results were properly documented in accordance with the test procedures. In
addition, the inspectors reviewed maintenance work records and corrective action
documents associated with the test rig and data acquisition system.
b. Description
From March 11-14, 2013, two NRC inspectors, along with one contractor, observed a
portion of Phase 1 and Phase 2 testing for squeeze film damping effects on a pair of
flat anti-vibration bars. Tests were conducted in air and water, with tube frequency
being varied. The test rig did allow for the adjustment of the simulated anti-vibration
bar. The tube was excited by a pair of electromagnetic coils with three eddy current
proximity probes used to measure tube vibration.
For a more detailed description of the inspection and testing performed at Chalk
River, refer to Attachments 3 and 4.
- 41 -
c. Findings
No findings were identified.
The inspectors did concur with the testing results that showed squeeze film damping
was negligible and did not contribute to the damping values assumed in the
operational assessments. At the time of the plant permanent shutdown
announcement, Southern California Edison had not completed revision of their return
to service plan operational assessments to account for these results. Therefore, the
inspectors did not complete an inspection of the affect of these results on the return to
service plan.
4OA6 Meetings
Exit Meeting Summary
On August 28, 2013, the inspectors presented the inspection results to Mr. P. Dietrich,
Senior Vice President and Chief Nuclear Officer, and other members of the licensee staff.
The licensee acknowledged the issues presented. Proprietary information was provided
to the inspectors.
Attachment 1
SUPPLEMENTAL INFORMATION
KEY POINTS OF CONTACT
Licensee Personnel
P. Dietrich, Senior Vice President and Chief Nuclear Officer
D. Bauder, Site Vice President
T. Palmisano, Vice President of Engineering, Projects and Site Support
B. Sholler, Director, Maintenance
O. Flores, Director, Nuclear Oversight
R. St. Onge, Director, Regulatory Affairs/Emergency Planning
E. Avella, Director, Project Management Organization
R. Davis, Director, Training
J. Madigan, Director, Safety Culture
C. McAndrews, Director, Special Projects
R.Treadway, Manager, Nuclear Regulatory Affairs
A. Martinez, Manager, Chemistry
L. Mosher, Manager, Communications
K. Yhip, Technical Advisor
M. Brown, Project Manager, Communications
R. Swanson, Consultant
K. Gallion, Manager, Focus Assessments/Performance Improvements
M. Pawlaczyk, Technical Specialist, Regulatory
NRC Personnel
D. Beaulieu, NRR, Project Manager
LIST OF ITEMS OPENED, CLOSED, AND DISCUSSED
Opened and Closed 05000361/2012009-01
Failure to Verify Adequacy of Thermal-Hydraulic and Flow-
Induced Vibration Design for the Unit 2 Replacement Steam
Generators
Opened 05000362/2012009-02
Failure to Verify Adequacy of Thermal-Hydraulic and Flow-
Induced Vibration Design for the Unit 3 Replacement Steam
Generators
Closed
Unit 3 Steam Generator Tube Degradation Indicated by Failed
In-Situ Pressure Testing
A1-2 05000362/2012007-04
Evaluation of Changes in Dimensional Controls during the
Fabrication of Units 2 and 3 Replacement Steam
Generators05000362/2012007-08
Nonconservative Thermal-Hydraulic Model Results
LIST OF DOCUMENTS REVIEWED
DESIGN BASIS
DOCUMENT
TITLE
REVISION
DBD-SO23-360
10, 11
DRAWING
TITLE
REVISION
L5-04FU001
Design Drawing - Component and Outline Drawing
6
L5-04FU011
Design Drawing - Channel Head 1/4
14
L5-04FU016
Design Drawing - Divider Plate 1/2
12
L5-04FU021
Design Drawing - Tubesheet and Extension Ring 1/3
16
L5-04FU031
Design Drawing - Lower Shell, Middle Shell, Transition Cone, and
Upper Shell 1/4
6
L5-04FU041
Design Drawing - Upper Shell, Upper Head Ring, and Upper Head
Top 1/4
6
L5-04FU051
Design Drawing - Tube Bundle 1/3
1
L5-04FU052
Design Drawing -Tube Bundle 2/3
1
L5-04FU053
Design Drawing - Tube Bundle 3/3
3
L5-04FU054
Design Drawing -Tubing Expansion and Seal Welding
2
L5-04FU101
Design Drawing - Wrapper Assembly 1/5
5
L5-04FU106
Design Drawing -Tube Support Plate Assembly 1/3
3
L5-04FU107
Design Drawing -Tube Support Plate Assembly 2/3
3
L5-04FU108
Design Drawing - Tube Support Plate Assembly 3/3
3
L5-04FX001
Fabrication Drawing, General Shipping Arrangement [SON-2A
(2E089), SON-2B (2E088)]
4
L5-04FX002
Fabrication Drawing, General Shipping Arrangement [SON-3A
(3E089), SON-3B (3E088)]
5
A1-3
DOCUMENT
TITLE
REVISION/DATE
UGNR-SON2-RSG-
067(7)
Nonconformance Report: 2563901/G101,
Unacceptable gaps between Tubes and AVBs.
January 23, 2007
UGNR-SON3-RSG-030
Mitsubishi Nonconformance Report - Some Gaps
between Tubes and AVBs are larger than the
criterion
0
SON-3A(3E089)-1, -2,
-3, -4
[Sumitomo] G-Values
September and
October 2007
SON-3B(3E088)-1, -2
[Sumitomo] G-Values
December 2007
SON-2A(2E089)-1, -2,
-3,
[Sumitomo] G-Values
October and
November 2006
SON-2B(2E088)
[Sumitomo] G-Values
December 2006
KAS-20130179
SONGS U-tube damping measurement test for
comparing with the test result of AECL: Test plan
0
KAS-20040251
FIVATS (Fluid Elastic Vibration Analysis Code)
Code Description Note (Users Manual)
1
KAS-20040252
FIVATS Code Validation and Qualification Plan
3
KAS-20040253
FIVATS Code Validation and Qualification Report
3,4
N/A
Damping Measurement Test Procedure
1
N/A
Sample Calculation of Stability Ratio
February 1, 2013
N/A
Evaluation of Liquid Film Thickness of Tube at AVB
Support Point
January 23, 2013
N/A
The procedure how frequencies are computed by
FIVATS
January 17, 2013
L5-04GA571
Screening Criteria for Susceptibility to In-Plane
Tube Motion
4
L5-04GA564
Tube Wear of Unit-3 RSG - Technical Evaluation
Report
2
L5-04GA428
Design of Anti-Vibration Bar
5
L5-04GA504
Evaluation of Tube Vibration
3,4
L5-04GA102
Nitrogen Plenum / Accelerometer Data Report for
Unit 3
1
L5-04GA224
Material Selection Report for Anti-Vibration Bar
2
L5-04GA521
Three-Dimensional Thermal and Hydraulic
Analysis (FIT-III Code Analysis)
3
A1-4
DOCUMENT
TITLE
REVISION/DATE
L5-04GA591
Validity of Use of the FIT-III Results during Design
1,3
UES-2010254
Root Cause Analysis Report for tube wear
identified in the Units 2 and 3 Steam Generators of
San Onofre Nuclear Generating Station
0
N/A
SGTL Unit 2 Mode 4 Entry JITT [S/G Tube Leak
Unit 2 Mode 4 Entry Just in Time Training]
December 2012
N/A
JITT for SGTR QUIZ [Just in Time Training for S/G
Tube Rupture Quiz]
December 2012
N/A
Plant Changes - Unit 2 Return to Service
information package, Lesson Plan 2RP548,
Attachment 9.2
December 2012
N/A
Scenario Title: Mode 1 and 2 SGTL and SGTR
May 18, 2012
N/A
SCE Purchase Order 4500555142 with CANDU
Energy, Inc.
January 28, 2013
N/A
SCE Change Order 1 to PO 4500555142
March 7, 2013
N/A
SONGS NOD memo to file, SUBJECT: LEAD
AUDITOR ANNUAL EVALUATION
December 11,
2012
N/A
Commercial grade item survey report, Survey
Report AECL-CS1-13
February 18-22,
2013
N/A
Letter to Elmo E. Collins from Southern California
Edison; Docket 50-361, Confirmatory Action Letter
- Actions to Address Steam Generator Tube
Degradation, San Onofre Nuclear Generation
Station Unit 2
October 3, 2012
N/A
San Onofre Nuclear Generating Station Unit 2
Return to Service Report
October 3, 2012
N/A
SONGS U2C17 Steam Generator Operational
Assessment
October 3, 2012
1814-AU651-MO157
SONGS Unit 2 Cycle 17 Steam Generator
Operational Assessment (AREVA)
0
1814-AU651-MO160
SONGS Unit 2 Cycle 17 Steam Generator
Operational Assessment for Tube-to-Tube Wear
(AREVA)
0
1814-AU651-MO145
Operational Assessment for SONGS Unit 2 SG for
Tube-to-Tube Wear Degradation at the End of
Cycle 16 (Intertek)
1
A1-5
DOCUMENT
TITLE
REVISION/DATE
1814-AA086-M0190
Operational Assessment of Wear Indications in the
U-Bend Region of San Onofre Nuclear Generating
Station Unit 2 Replacement Steam Generators
4
SO23-617-01
Specification for Design and Fabrication of the
Replacement Steam Generators for Unit 2 and
Unit 3
4
1814-AV651-M0165
SONGS Unit 2, Steam Generator Internal Impact
Test Results
0
1814-AU651-MO151
SONGS U2C17 and U3F16B AVB Gap and Tube-
to-Tube Proximity Measurement Program
0
LTR-SGDA-12-36
Flow-induced Vibration and Tube Wear Analysis of
the San Onofre Nuclear Generating Station Unit 2
Replacement Steam Generators Supporting
Restart
1
SG-SGMP-12-10
Operational Assessment of Wear Indications in the
U-bend Region of San Onofre Nuclear Generating
Station Unit 2 Replacement Steam Generators
3
NECP 800873488-132
50.59 Screen 70% Power Operation
December 19,
2012
NECP 800873488-124
50.59 Evaluation 70% Power Operation
October 15, 2012
NECP 800901029-
0041
50.59 Screen 70% Power Annunciator
September 14,
2012
NECP 800867185-
0021
50.59 Screen Argon Injection into RCS
November 19,
2012
NECP 800905312-
0270
50.59 Screen Nitrogen-16 Monitors
December 3, 2012
NECP 800698429-
0140
50.59 Screen U2C17 Reload ECP
August 30, 2012
NECP 800698429-
0150
50.59 Screen for TR-PL Correction
July 30, 2012
NECP 800873488-
0130
50.59 Screen Plugging and Stabilization
December 6, 2012
NECP 800457837-
0550
50.59 Screen Vibration and Loose Parts Monitoring
System
January 14, 2013
A1-6
DOCUMENT
TITLE
REVISION/DATE
NN 201843216
RCE Steam Generator Tube Wear, San Onofre
Nuclear Generating Station, Unit 2
April 23, 2012
NECP 800905312-
0270
50.59 Screen Nitrogen-16 Monitors
December 3, 2012
NECP 800698429-
0140
50.59 Screen U2C17 Reload ECP
August 30, 2012
MISCELLANEOUS
DOCUMENT
TITLE
REVISION/DATE
Standards of The Tubular Exchanger Manufacturer
Associations (TEMA), 8th Edition
1999
N/A
Steam Generator RTS Schedule
AECL Testing
February 19, 2013
1814-AD799-M0001
AECL EACL Test Plan, Measurement of Steam-
Generator Damping Due to Anti-Vibration Bar
Supports
1
51-9198780-000
[AREVA] Distribution of Indications for SONGS 2
and 3 - 2012 Inspections
February 27, 2013
51-9182205-000
[AREVA] SONGS Unit 3 2012 Forced Outage
(U3F16B)
Technical Summary Steam Generator Eddy Current
Inspection
May 23, 2012
51- 9188725-001
[AREVA] SONGS U2C17 and U3F16B AVB Gap
and Tube-to-Tube Proximity Measurement Program
November 16,
2012
N/A
AECL Damping Tests for SONGS (PowerPoint
Presentation)
March 8, 2013
Proceedings of the
Institution of
Mechanical Engineers,
Volume 184, Part 1,
No. 36
Void Fractions In Two-Phase Flow: A Correlation
Based Upon An Equal Velocity Head Model
1969
Journal of Mechanical
Design, Volume 100
Fluidelastic Vibration of Heat Exchanger Tube
Arrays
April 1978
10th International
Conference on Flow-
Induced Vibration (&
Flow-Induced Noise)
Study on In-flow Fluid-elastic Instability of Circular
Cylinder Arrays
2012
A1-7
MISCELLANEOUS
DOCUMENT
TITLE
REVISION/DATE
Journal of Vibration,
Acoustics, Stress, and
Reliability in Design,
Volume 105
The Effect of Approach of Flow Direction on the
Flow-Induced Vibration of a Triangular Tube Array
January 1983
Journal of Pressure
Vessel Technology,
Volume 126
Damping of Heat Exchanger Tubes in Two-Phase
Flow: Review and Design Guidelines
November 2004
Journal of Pressure
Vessel Technology,
Volume 128
Fluidelastic Instability of an Array of Tubes
Preferentially Flexible in the Flow Direction
Subjected to Two-Phase Cross Flow
February 2006
Journal of Pressure
Vessel Technology,
Volume 127
Fluidelastic Instability and Work-Rate
Measurements of Steam-Generator U-Tubes in Air-
Water Cross-Flow
February 2005
IMECE2002-32707
Vibration Analysis of Steam Generators and Heat
Exchangers: An Overview Part 1: Flow, Damping,
Fluidelastic Instability
November 2002
Journal of Applied
Mechanics, Volume 47
Fluid Forces on Rods Vibrating in Finite Length
Annular Regions
June 1980
Journal of Pressure
Vessel Technology,
Volume 133
Damping of Heat Exchanger Tubes in Liquids:
Review and Design Guidelines
February 2011
Journal of Engineering
for Power,
Volume 105
The Effect of Flat Bar Supports on the Crossflow
Induced Response of Heat Exchanger U-Tubes
October 1983
Nuclear Technology,
Volume 55
Flow-Induced Vibration and Wear of Steam
Generator Tubes
November 1981
Journal of Pressure
Vessel Technology,
Volume 117
Vibration of a Tube Bundle in Two-Phase Freon
Cross-Flow
November 1995
Nuclear Insights
Fluid Elastic Instability Causing Tube Damage in
Main Steam Condensers of Nuclear Power Plants
Spring 2009
Nuclear Engineering
and Technology,
Volume 38, No. 1
Fluid-Elastic Instability of Rotated Square Tube
Array in an Air-Water Two-Phase Cross-Flow
February 2006
R&D 13-2259
Rust on AVB material
March 12, 2013
R&D 13-2279
Free span length 2.993 m vs. 3.000 m in test plan
March 12, 2013
A1-8
MISCELLANEOUS
DOCUMENT
TITLE
REVISION/DATE
R&D 13-2278
Temperature recording - air temperature not
recorded
March 12, 2013
R&D 13-2277
Circular hole support plate thickness
March 12, 2013
R&D 13-2372
Force transducer
March 14, 2013
R&D 153-127370-TP-
001
AECL/EACL Test Plan, Measurement of Steam
Generator Tube Dampening Due to Antivibration
Bar Supports
D1, D2, and O
CW-510200-PRO-344
AECL/EACL procedure: Assessment and
Identification of Personnel Qualifications and
Training Needs
2
CW-510100-FM-164
AECL/EACL Position Description, Senior R&D
Engineer
4
N/A
AECL/EACL student training history, TRAK printout
(4 students/employees)
1999 to 2013
CW-510200-PRO-498
AECL/EACL employee training, Approving,
Attending, and Reading
1
CW-510200-FM-240
AECL/EACL Personnel Qualifications and Training
Needs (4 students)
4
N/A
AECL/EACL Technical Document 0987-00, Test
Instruction for Straight-Tube Dampening Tests for
Southern California Edison
0, 5
N/A
AECL/EACL Technical Document 0987-00, Test
Preparation Instructions for Straight-Tube
0, 5
153-127370-TR-002
[AECL] Measurement of Steam-Generator Tube
Damping Due to Anti-Vibration Bar Supports
0
1022832
Steam Generator Management Program: PWR
Primary-to-Secondary Leak Guidelines
4
TR-016743-V2R1
EPRI Guidelines for PWR Steam Generator Tubing
Specifications and Repair, Volume 2
1
TR-103824s-V1R1
EPRI Steam Generator Reference Book, Volume 1
1
LTR-SGMP-12-70
Distributions of Gap Velocities, Void Fractions and
Fluid Densities along Selected Tubes of the
SONGS Replacement Steam Generators
December 4, 2012
12 -9176741-002
[AREVA] Technical Data Record for Stabilizer
Design
April 27, 2012
A1-9
MISCELLANEOUS
DOCUMENT
TITLE
REVISION/DATE
12 -9176741-005
[AREVA] Technical Data Record for Stabilizer
Design
October 16, 2012
N/A
Advisory Committee on Reactor Safeguards
Materials, Metallurgy and Reactor Fuels Steam
Generator Action Plan
September 24,
2009
A-SONGS-9416-1168
Evaluation of Southern California Edison Songs
Unit 3 Steam Generators with Degraded Eggcrates
1
LTR-SGDA-12-50
ATHOS Computer Code Verification & Validation
Summary Report for SONGS Unit 2 RSG Recovery
Project
September 14,
2012
90202
Comparison Of FIV Structural Models
0
NUCLEAR NOTIFICATIONS
200525719
201937413
202243314
201836127
201969741
201836127
201279476
201969131
201907105
201843216
201979105
201843216
PROCEDURE
TITLE
REVISION
SO23-12-4
Steam Generator Tube Rupture
23, 24
SO23-5-1.7
Power Operations
54
SO23-3-2.21
Core Operating Limits Supervisory Limits (COLSS)
29
SO23-15-50.A2
ARP Annunciator Panel 50A, PZR/CEA Window 31-60
19
SO23-XXXVII-1.20
RCS Calorimetric Flow Measurements
6
SO23-3-3.25
Once a Shift Surveillance (Mode 1-4)
37
SO23-3-2.1
Operation of Pressurizer Degas System
39
SO23-15-50.A1
ARP Annunciator Panel 50A
13
SO123-III-2.22.23
Unit 2/3 Steam Generator Tube Leakage Monitoring
Program
25
SO23-13-14
Reactor Coolant Leak
16, 17, 21
SO23-3-2.1.1
CVCS Alignment
16, 18, 20
SO23-3-2.24 ISS 2
Radiation Monitoring System Guidelines and RDU
Operation
14
SO23-3-2.1
CVCS Operation
40
A1-10
PROCEDURE
TITLE
REVISION
SO123-III-1.1.23
Unit 2/3 Chemical Control of Primary Plant and Related
Systems
59, 60
SO23-3-3.37
Reactor Coolant System Water Inventory Balance
35
SO123-XII-2.19
Qualification and Certification of Auditing Personnel
8
SO123-XII-2.20
Certification Documentation Processing for Personnel
Involved in Audits, Inspections, and Examinations
10
Attachment 2
Independent Evaluation Of San Onofre Nuclear Generating Station (SONGS) Steam Generator
Tube Wear Problems
Ian Barnes, Beckman & Associates
Hans Garkisch, Beckman & Associates
July 13, 2012
1. INTRODUCTION
An Augmented Inspection Team (AIT) was approved by the Nuclear Regulatory Commission
(NRC) on March 16, 2012, to assess the facts and circumstances surrounding the occurrence of
a tube leak and detection of unexpected tube wear in SCE Unit 3 steam generators (SGs). The
AIT was established in accordance with NRC Management Directive 8.3, NRC Incident
Investigation Program, and implemented using Inspection Procedure 93800, Augmented
Inspection Team.
In addition to performance of the AIT inspection, the NRC requested performance of this
independent evaluation, with a primary focus to be identification of any perceived gaps in the
response actions taken by the NRC or licensee.
2. BACKGROUND, EVENT DESCRIPTION, AND INSPECTION RESPONSE
Replacement steam generators for SCE Units 2 and 3 were designed and fabricated by
Mitsubishi (MHI), with Unit 2 return to power after installation occurring on April 13, 2010 and
Unit 3 on February 18, 2011. The replacement steam generators were designed to be like-for-
like replacements for the original steam generators, but included major improvements such as:
(a) replacement of carbon steel egg crate design tube support plates (tube support plates) with
Type 405 ferritic stainless steel tube support plates using trefoil hole configurations, and (b) use
of the significantly more corrosion resistant thermally treated Alloy 690 tubing in place of the
original mill annealed Alloy 600 tubing.
Unit 2 was shut down for a scheduled refueling outage on January 10, 2012. During the
required inservice inspection (ISI) activities, unexpected steam generator tube wear was
detected. Specifically, pluggable tube wear was found in: (a) two steam generator 2E0-89
tubes that had been caused by retainer bars, and (b) four steam generator 2E0-88 tubes that
had been caused by anti-vibration bars (anti-vibration bars) for two tubes and by retainer bars
for two tubes. As a result of the unexpected wear, the licensee preventatively plugged an
additional 92 tubes in steam generator 2E0-89 and an additional 94 tubes in steam generator
During the Unit 2 refueling outage, the Unit 3 control room received an alarm on January 31,
2012 from the main condenser air ejector radiation monitors that indicated a primary-to-
secondary tube leak in steam generator 3E0-88. The initial estimated leak rate was 75 gallons
per day (gpd) and was increasing, versus a facility license requirement of less than 150 gpd
steady state leak rate. Shutdown was initiated with cold shutdown reached on February 2,
2012. Subsequent inspection identified that the location of the leak in steam generator 3E0-88
was the tube in the Row 106 Column 78 (R106C78) location, with no other tubes found to be
leaking. Subsequent eddy current inspections of all of the tubes in the Unit 3 steam generators
discovered unexpected wear in both replacement steam generators, including significant tube-
to-tube wear in the freespan areas of the U-bends and tube wear at retainer bars similar to that
A2-2
identified in the Unit 2 steam generators. Condition monitoring requirements were not met for
129 tubes (73 tubes in steam generator 3E0-88, 56 tubes in steam generator 3E0-89), which
required in-situ pressure testing of the tubes. Eight steam generator 3E0-88 tubes failed the in-
situ pressure test. As a result of the wear findings in the Unit 3 steam generators, additional
eddy current testing inspections were performed of the Unit 2 steam generators, which included
use of the more sensitive +Point rotating probe. The area included all rows above R79 between
columns70-110.
Table 1 shows summary information regarding the extent of measurable tube wear that was
detected by eddy current testing inspections of the Unit 2 and Unit 3 steam generators. The
most significant difference between the units is the extent of detected tube-to-tube wear (tube-
to-tube wear) in the Unit 3 steam generators versus that found in the Unit 2 steam generators.
Of the 326 affected Unit 3 steam generator tubes exhibiting tube-to-tube wear, eddy current
testing inspection identified a total of 202 to have pluggable 35 percent through-wall depth
(TWD). Neither of the two affected Unit 2 steam generator tubes showing tube-to-tube wear
were found to have 35 percent TWD, but were preventatively plugged and stabilized. A
significant incidence of tube wear was detected in both units at the anti-vibration bars and tube
support plates. From review of Root Cause Evaluation (RCE) 201843216, it was ascertained
that eddy current testing inspection identified at tube support plate locations a total of 230 tubes
in the Unit 3 steam generators with pluggable 35 percent TWD versus 0 tubes with
35percent TWD in the Unit 2 steam generators. The Unit 3 steam generator tube degradation
occurred in slightly less than one year of power operations compared to a full cycle for Unit 2.
Table 1
Steam Generator Extent of Condition
Tube Degradation
Type/Location
Number of Unit 2 Affected
Tubes
Number of Unit 3 Affected
Tubes
Tube-to-Tube Wear
2
326
Wear at Anti-Vibration
Bars
1399
1767
Wear at Tube Support
Plates
299
463
Wear at Retainer Bars
6
4
Wear at Foreign Objects
2
0
Unit 2 and Unit 3 steam generators were found by eddy current testing inspection to both
contain a total of two tubes with pluggable 35 percent TWD at anti-vibration bar locations.
Eddy current testing inspection wear depth measurements with reasonable credibility (i.e.,
20 percent TWD or greater) did not indicate any particular difference between Unit 2 and Unit 3
steam generators regarding tube wear at anti-vibration bars. Specifically, the respective totals
for tubes with an eddy current testing measured depth of 21-30 percent TWD at anti-vibration
bar locations were 52 for steam generator 2E0-88, 51 for steam generator 2E0-89, 19 for steam
generator 3E0-88, and 31 for steam generator 3E0-89. Taking into consideration the difference
in operating cycle time between Units 2 and Unit 3, these numbers suggest little difference in
tube wear behavior between Units 2 and 3 steam generators at anti-vibration bar locations.
This behavior was considered somewhat surprising, considering the extent of Unit 3 steam
generator freespan tube-to-tube wear and the apparent resulting enhanced tube wear at upper
tube support plates that occurred from transfer of this vibration energy.
A2-3
What is clear from reviewing inspection data for the steam generators is that wear for both
Units 2 and 3 steam generators occurred in the same localized region of the tube bundle, which
suggests a common thermal-hydraulic root cause.
3. SUMMARY OBSERVATIONS
No evidence was found in the reviewed documentation that addressed the cause for the
abnormal and localized high void fraction in the replacement steam generator design, which
historical degradation information would indicate was absent in the original steam generator
design. It would also appear that the existence of the localized high void fraction and flow
velocities, as calculated by ATHOS, was not questioned as a replacement steam generator
design feature or compared against the replacement steam generator design basis. Rather,
the thermal-hydraulic analysis results were accepted, initially by design, and again during
the extensive review process.
The conditions leading to the wear were calculated with the EPRI ATHOS code, a standard
steam generator design tool. The calculated high void fraction region included the smaller
region of tubes with the observed wear. It was not obvious that a small region with
conditions which caused tube wear can be extracted from the ATHOS results.
Any conclusions to be drawn from contribution of tube ovality to tube wear differences
between Unit 2 and Unit 3 steam generators should specifically consider the Unit 2 and
Unit 3 G-Values of tubes in the affected Unit 3 sub-population wear region. The tube ovality
decreases gaps between tubes and anti-vibration bars, but also increases the propensity for
in-plane vibration due to the decrease of cross-section stiffness.
Absent the existence of additional information, there is no apparent basis to believe that the
number of local radius adjustments during manufacture of U-bends has any relevance to
observed steam generator tube degradation.
The average of the gaps between the outermost tubes and the central columns was found to
be essentially the same between the Unit 2 and Unit 3 steam generators, which does not
support a premise that more uniform manufacturing practices for Unit 3 steam generator
tube bundles resulted in less contact force between anti-vibration bars and tubes. In the
absence of more dimensional information for the steam generator tube bundles, it is not
believed possible to explicitly define the number of active supports in the Unit 2 and Unit 3
Eddy current testing inspection measurements of tube-to-anti-vibration bar gap were
determined to be of questionable value in an assessment of likely tube wear behavior.
The performance conditions calculated with the thermal-hydraulic codes ATHOS and FIT-III
are inconsistent with the thermal-hydraulic design described in L5-04GA510, Rev. 5,
Thermal and Hydraulic Parametric Calculations. The consequences of this inconsistency
between the replacement steam generator system design and calculated thermal-hydraulic
performance has not been addressed.
No evidence was found that the adoption of a smaller tube pitch/tube diameter ratio of 1.33
compared with the original steam generator, with a potential choking effect on inlet
secondary flow (from the wrapper ports), was evaluated in design or addressed in reviews.
A2-4
4. PROBABLE CAUSE EVALUATION
Condition Report NN 201836127, Root Cause Evaluation: Unit 3 Steam Generator Tube Leak
and Tube-to-Tube Wear Condition, Revision 0, identified the mechanistic cause of tube-to-tube
wear in SGs 3E0-88 and 3E0-89 to be fluid elastic instability (FEI) involving the combination of
localized high steam/water velocity (tube vibration excitation forces), high steam void fraction
(loss of ability to dampen vibration), and insufficient tube-to-anti-vibration bar contact forces to
overcome the excitation forces. The results of and visual inspections strongly support FEI being
the applicable tube-to-tube wear damage mechanism. Section 8 of the draft AIT report
concluded from independent NRC ATHOS code thermal-hydraulic analysis that the SCE
replacement steam generators were not designed with adequate margin to preclude onset of
FEI. The AIT inspection also concluded that the deficiencies appear to be related to the
Mitsubishi FTI-III thermal-hydraulic code having predicted non-conservative low velocity results.
Use by Mitsubishi of the ATHOS code and independent thermal-hydraulic analyses by
Westinghouse (using a company version of the EPRI ATHOS code) and by AREVA (using a
proprietary code) arrived at similar velocity conclusions as reached by the independent NRC
ATHOS code review.
It was concluded from this review that the fluid condition alone did not explain the abnormal
tube-to-tube wear based on the standard FEI criteria of ASME Boiler and Pressure Vessel
Code,Section III, Division1, Appendix N, Section N-1330, 1998 . As noted in both the AIT draft
report and the SCE RCE, no tube wear of the type detected in the Unit 3 replacement steam
generators has been previously observed in other recirculating SGs in the domestic fleet. The
observed severe tube-to-tube wear was also restricted to a small region of the tube bundle
cross section. Based on the wear indications, the U-bend tube support by the anti-vibration
bars was ineffective in a small region of the tube bundle for in-plane tube vibrations. Various
assumptions were made to rationalize the mechanism, which caused the observed tube-to-tube
wear including: (a) insufficient tube support at multiple anti-vibration bar locations, (b) vibration
in an in-plane mode, (c) gaps between tubes and anti-vibration bars, and (d) spreading of the
upper U-bend structure due to fluid dynamic forces and thermal effects.
No evidence was found in the reviewed documentation that was pertinent to the following:
(a) the cause for the abnormal and localized high void fraction in the replacement steam
generator design, which historical degradation information would indicate was absent in the
original steam generator design; and (b) the existence of the localized high void fraction and
flow velocities, as calculated by ATHOS, and originally by FIT-III with lower flow velocities, that
has apparently been accepted without question as a replacement steam generator design
feature. No evidence was found that the flow conditions, void fractions, flow velocities and
temperatures of the tube bundle were compared against the replacement steam generator
design basis. Rather, the thermal-hydraulic analysis results were accepted, initially by design,
and again during the extensive review process.
The conditions leading to the wear were calculated with the EPRI ATHOS code, a standard
steam generator design tool. The calculated high void fraction region included the smaller
region of tubes with the observed wear. It was not obvious that a small region with conditions
which caused tube wear can be extracted from the ATHOS results.
The U-bend tube support conditions were evaluated in detail by Mitsubishi and in the root cause
evaluation, and a preliminary status of these evaluations is described in the draft AIT inspection
report. The anti-vibration bar and tube tolerances were considered in estimating the tube
A2-5
support; however, the component and design tolerances, as described in Document L5-
04GA428, Design of Anti-Vibration Bar, Revision 5, were not stacked across anti-vibration
bars, tubes, and the tube bundle to accumulate maximum possible gaps or interferences, and
determine the potential extremes of the tube support conditions that would be appropriate for a
loose bundle.
5. DESIGN AND MANUFACTURING DIFFERENCES
The AIT inspection did not identify any significant differences in the design requirements of the
Unit 2 and Unit 3 replacement steam generators. This evaluation also did not note any
significant differences in design requirements between the Unit 2 and Unit 3 replacement steam
generators.
The AIT inspection identified two unresolved items during its review pertaining to: (a) retainer
bar-to-tube wear in the Unit 2 replacement steam generators, and (b) consideration of the
potential impact of improving dimensional controls for tube roundness and anti-vibration bars. It
was concluded from review of these unresolved items that the retainer bar-to- tube wear issue
warranted no additional comment. Review of the subject material pertaining to dimensional
controls of tube roundness and anti-vibration bars led to the following observations:
The standard deviation for tube O.D. in the Unit 2 and Unit 3 replacement steam generators
was calculated by Mitsubishi to be: steam generator 2E0-88, 0.71 mils; steam generator
2E0-89, 0.71 mils; steam generator 3E0-88, 0.63 mils; steam generator 3E0-89, 0.55 mils.
Mitsubishi postulated in Report L5-04GA564, Rev. 2 that improved dimensional controls for
Unit 3 replacement steam generators such as anti-vibration bar thickness, tube roundness,
and gaps between tubes and anti-vibration bars probably resulted in less contact force
between the tubes and the anti-vibration bars. This difference in standard deviation values
for the O.D. of the tube populations in the individual SGs is considered by this review to
have minimal effect, particularly if one takes into consideration that the low radius U-bends
are the biggest contributor to tube ovality and higher G-values. The localized region of tube
wear, however, is located in high row number tubes where variations in G-values would not
be expected in the large radius U-bends during ongoing production. It is believed that any
conclusions to be drawn from contribution of tube ovality to wear differences between Unit 2
and Unit 3 SGs should specifically consider the Unit 2 and Unit 3 G-Values of tubes in the
Unit 3 sub-population wear region.
Mitsubishi Report L5-04GA564, Revision 2, noted that the number of adjustments to tube
bending radius was smaller for the Unit 3 SGs than for the Unit 2 SGs. Specifically, the
reported values were: steam generator 2E0-88, 265; steam generator 2E0-89, 390; steam
generator 3E0-88, 132; steam generator 3E0-89, 149. The inference drawn was in the
context of promoting greater uniformity in tube to anti-vibration bar gaps. The required
profile for U-bends is established on an inspection layout table and needed local, minor
radius adjustments are made to assure conformance to the required profile. Absent the
existence of additional information, there is no apparent basis to believe that the number of
local adjustments to U-bends has any relevance to observed steam generator tube
degradation.
The average gap between outermost tubes and anti-vibration bars for the Unit 2 and Unit 3
SGs was reported by Mitsubishi Report L5-04GA564, Rev. 2 to be: steam generator 2E0-
88, 0.59 mils; steam generator 2E0-89, 0.76 mils; steam generator 3E0-88, 0.15 mils; steam
generator 3E0-89, 0.21 mils. This report additionally stated that the average of the gaps
A2-6
between the outermost tubes and the central columns is essentially the same between the
Unit 2 and Unit 3 SGs. This data obviously does not support the premise that more uniform
manufacturing practices for Unit 3 steam generator tube bundles resulted in less contact
force between anti-vibration bars and tubes. In the absence of more dimensional
information for the steam generator tube bundles, it is not believed possible to explicitly
define the number of active supports in the Unit 2 and Unit 3 SGs.
Data was not specifically searched for during this review to allow formal assessment of the
technical credibility of eddy current testing inspection for measurement of gaps between
tubes and anti-vibration bars; i.e., it is currently unknown whether a qualified Examination
Technique Specification Sheet (ETSS) exists for this measurement. Difficulties in use of the
CERTREC system for information retrieval negatively affected conduct of this review.
Paragraph 4.1.2 of Mitsubishi Report L5-04GA564, Rev. 2 states, in part, with respect to
comparison of bobbin probe signals in the Unit 2 and Unit 3 SGs for estimating tube-to-anti-
vibration bar gap sizes This data did not reveal significant differences and indicates that
the gaps in the affected region of the tube bundle are below 20 mils (0.5 mm). However, the
average voltage signal in the Unit 3 SGs is slightly lower than the average signal in the Unit
2 SGs, indicating that the average gap size in the Unit 3 SGs is slightly larger than in the
Unit 2 SGs, and indicating that the average contact force between the tubes and anti-
vibration bars during operation may be lower in the Unit 3 SGs. These comments are
believed to be speculative, and rely on a global number of unknown technical credibility for
predicting values in a bundle sub-population. Review of Figure 4.1.2-1 in Report L5-
04GA564, Rev. 2 indicates the potential fallacy in making these projections. Specifically,
Figure 4.1.2-1shows virtually identical average absolute signal amplitude signals at anti-
vibration bar locations for SGs 2E0-88 and 3E0-89, SGs that have shown significant
differences in operational tube wear behavior. Accordingly, this review concluded that eddy
current testing inspection measurements of tube-to-anti-vibration bar gap were of
questionable value in assessment of likely tube wear behavior.
The most significant fabrication difference between the Unit 2 and Unit 3 replacement steam
generators relates to the cracking indications that were identified in both Unit 3 replacement
steam generators in the weld between the divider plate and the channel head subsequent to
completion of the ASME Section III Code primary side hydrostatic test. Repair of these defects
necessitated removal of the channel head from each of the Unit 3 steam generators. As a
result, the tubesheet-to-channel head circumferential weld had to be repeated and further post
weld heat treatment (PWHT) and hydrostatic tests performed. Review of Section 12.0 of the
draft AIT report found that the team had comprehensively reviewed divider plate repair activities.
An unresolved item was identified by the team pertaining to the adequacy of Mitsubishi
evaluation and controls for the divider plate weld repairs. Subject areas in question included:
(a) the approximate 300 additional rotations of each Unit 3 replacement steam generator that
resulted from additional welding of the channel head to tubesheet, and the lack of consideration
by Mitsubishi of the potential impact of these rotations on tube bundle configuration in terms of
anti-vibration bar gaps or distortion; (b) the lack of a full assessment by Mitsubishi of the impact
of heat input activities such as local PWHT, grinding and flame cutting on steam generator
configuration in terms of tubesheet thermal expansion or distortion; and (c) the non-performance
by Mitsubishi of dimensional checks after repair to confirm that critical secondary side
dimensions were not affected by the repairs.
This review concluded that local post weld heat treatment of the channel head to tubesheet
weld had the highest potential (of the AIT noted activities) for affecting the tube bundle. This
view derives from the possible effects of the temperature gradient across the tubesheet that is
A2-7
created by the local post weld heat treatment cycle applied to the channel head-to-tubesheet
weld. The gradient creates progressively lower thermal expansion in the legs of U-bends as the
distance increases from the periphery to the center of the tubesheet. This variation in U-bend
thermal expansion has resulted in the past in the detection of tube ding (DNG) eddy current
testing indications at the upper tube support plate location in replacement steam generators. A
DNG sort was requested from SCE. Limited review of the supplied information did not,
however, identify any correlation of DNG signals with performance of local post weld heat
treatment cycles, or any noted incidence at tube support plate locations in the replacement
6. REVIEW OF REPLACEMENT STEAM GENERATOR THERMAL-HYDRAULIC
DOCUMENTATION
6.1 Scope
This independent review of the thermal /hydraulic related aspects of the SCE steam generator
condition was primarily limited to available documentation in the CERTREC system and focused
on the features and operating conditions, which caused or could have contributed to tube
damage. The review focused on the replacement steam generator tube bundle, anti-vibration
bar and retainer bar designs, and the tube bundle flow condition. No specific difference was
noted with respect to the basic findings of the AIT and the RCE reports. The high void fraction
in the U-bends was, however, viewed as abnormal and not intended by design. Accordingly, a
review was performed, which took into consideration: (a) inspection evidence, tube wear in the
U-bends and at the tube support plates; (b) results from the ATHOS, FIT-III, and FEI
calculations; (c) the replacement steam generator system thermal-hydraulic design; and (d) a
comparison of the replacement steam generator changes from original steam generator design
features. This review led to a sequence of conclusions, which trace the origin of the abnormal
event conditions to a potential source not addressed by either the AIT inspection or the RCE
reports.
6.2
Thermal-Hydraulic Overview
6.2.1 Sub-cooled Height Above the Tubesheet
Document L5-04GA510, Rev. 5, Thermal and Hydraulic Parametric Calculations, defined
generic parameters for the replacement steam generator design and operation. The
calculations were noted to be easy to follow and provided replacement steam generator system
flows, pressures and temperatures for a range of operating conditions, with input parameters
listed for the thermal-hydraulic analysis. Design requirement are specified in SO23-617-01,
Rev. 4, Specification for Design and Fabrication of replacement steam generators for Units 2 &
3. Specific requirements and acceptance criteria are listed by reference to other documents. A
spot-check of the design requirements and acceptance criteria showed that they were satisfied
by the calculated system parameters. The intended tube bundle flow management is indicated
in Figure 1 of Appendixes 10 and 11 to Document L5-04GA510, Rev. 5. These figures identified
a sub-cooled height, HPH, at the bottom of the tube bundle, which was calculated with SG
Steady State Performance Calculation Code (SSPC Code) to be HPH 1364.965 mm for a Thot
temperature of 598 0F and HPH 1145.173 mm for a Thot temperature of 611 0F. Accordingly,
boiling was not intended to start at the tube sheet level, or below tube support plate 1, as
calculated by the thermal-hydraulic codes in L5-04GA521, Rev. 3 and confirmed by independent
calculations with ATHOS. The performance conditions calculated with the thermal-hydraulic
codes are thus inconsistent with the thermal-hydraulic design described in L5-04GA510, Rev. 5.
A2-8
The consequences of this inconsistency have not been addressed. Additionally, no requirement
was found that this non-boiling level should be maintained.
6.2.2 Flow from the Wrapper Inlet Ports to the Tube Bundle
No documentation was found, which detailed the flow from the wrapper inlet ports to the bottom
of the tube bundle, other than the result of the thermal-hydraulic calculations that modeled the
flow conditions. The only apparent change of the secondary flow condition from the original
steam generator to the replacement steam generator design is the smaller pitch-to-diameter
ratio of the tube bundle in the replacement steam generator (i.e., P/D = 1.33 in the replacement
steam generator, P/D = 1.433 in the original steam generator) and 327 more tubes. All other
parameters for the like-for-like steam generator replacements are essentially unchanged. No
documentation was found for comparing the design of the wrapper inlet ports between the
replacement steam generator and the original steam generator due to CERTREC administration
problems. The smaller pitch-to-diameter ratio of the replacement steam generator tube bundle
increases the cross-flow resistance in the tube bundle. As a result, the penetration of the flow
from the wrapper ports into the tube bundle and the flow distribution in the bundle changes
when compared with the original steam generator. The inlet flow stagnated in a region outside
the outer row of stay rods, visible on the flow velocity plots of ATHOS and FIT-III. No evidence
was found that this change of the effect of tube bundle inlet flow distribution was evaluated in
design or addressed in reviews.
6.2.3 Thermal-hydraulic Analysis
6.2.3.1 Mitsubishi FIT III Code Analysis
Document L5-04GA521, Three Dimensional Thermal and Hydraulic Analysis (FIT III Code
Analysis) calculated the void fraction and flow velocities of the tube bundle in some detail. The
void fraction, flow velocities and temperatures throughout the tube bundle are presented and
are the input to the vibration calculations. Issues with this code were addressed in the draft AIT
inspection report. The AIT inspection noted that the Mitsubishi reported flow velocities and void
fractions appeared to be low and, as a result, performed independent calculations using the
EPRI ATHOS code. These calculations found much higher velocities in the tube bundle. The
cause for the FIT-III discrepancy has currently not been resolved. The AIT inspection also noted
that the validation and verification of the FIT-III code did not provide sufficient evidence that the
code had been adequately benchmarked. The AIT inspection concluded, without performing
vibration analysis, that the higher flow velocities and void fractions were the cause of the
observed FEI and tube wear.
6.2.3.2 ATHOS and AREVA CAFCA Analyses
After the identification of FIT III code issues, an ATHOS thermal-hydraulic analysis was
performed by the replacement steam generator supplier, Mitsubishi, with results compared
against the independent analysis results from the NRC (ATHOS), Westinghouse (Company
ATHOS version) and AREVA (CAFCA4 Code). The results of these calculations appeared to be
in reasonable agreement, with some variation of results due to modeling differences, and all
showed a local region of high void fractions starting below tube support plate 7 and extending
into the U-bend region. In normal steam generator design, one would expect a relatively flat
distribution of the void fraction level increasing from the sub-cooled region. The void fraction and
fluid velocities were noted to be high in a local U-bend tube region of the bundle where the
tubes exhibited significant wear.
A2-9
With ATHOS calculated flow conditions as input, the critical velocities and stability ratios
obtained with the Connors equation in the ASME Boiler and Pressure Vessel Code Section III,
Division1, Appendix N Section N-1330 still yielded stability ratios (ratio of effective flow to critical
FEI flow) <1 for tubes with effective restraint at tube support plate and anti-vibration bar support
locations. Accordingly, for FEI to occur, critical cross-flow velocities have been assumed to
occur for U-bend tube sections not restrained from in-plane motion by insufficient tube-to-anti-
vibration bar contact. No information was noted during documentation review regarding why
U-bend tubes with larger U-bend radius, outside the small tube-to-tube wear damage region
with high void fractions, did not experience the FEI tube movement, because with a larger
curvature they were more susceptible to tube vibration unless the restraints were more effective.
One observation from examining these calculation results was that the nucleate boiling started
in a region outside the outer row of stay rods. Figure 8.3-3 of Document L5-04GA521, Rev. 3
showed local void fractions, and Figures 8.3-1 and 8.3-2 showed the flow velocities above the
tubesheet. Nucleate boiling at the tubesheet surface is an anomaly not intended to occur in the
intended design as described in Mitsubishi Document L5-04GA510, Rev.5. The consequences
of this inconsistency between the replacement steam generator system design and calculated
thermal-hydraulic performance has not been addressed. It was also noted that Westinghouse in
their independent ATHOS analysis indicated that their design approach precluded boiling at the
tubesheet surface. The graphs of flow pattern above the tubesheet indicate a region of low flow
velocities where higher void fractions than in the surrounding fluid are indicated. A potential
cause for these low flow velocities, a region of almost stagnant flow, is the higher flow
resistance for the cross-flow from the wrapper inlet ports into the tube bundle due to the smaller
pitch-to-diameter ratio of the replacement steam generators than in the original steam
generators. A comparison of the replacement steam generator thermal-hydraulics with that of
the original steam generator was not found in either the AIT or the RCE reports, which could aid
in the determination of the cause for the flow abnormalities in the replacement steam generator.
6.2.4 Tube Wear at tube support plates above the Tube Sheet Hot-Spot
Mitsubishi Report L5-04GA564, Tube Wear of Unit-3 RSG - Technical Evaluation Report,
Revision 2, includes figures showing the distribution of tube wear at tube support plate 1 to 7
levels and in the U-bend sections. The region at the tube support plate 1 level with TWD tube
wear is immediately above the location where nucleate boiling at the tube sheet level started. A
reasonable supposition is that flow caused this local wear. The correlation between the tube
wear increasing upward from tube support plate to tube support plate and the location of the hot
spot appeared to be systematic in the replacement steam generators.
The coolant flow between tubes in the straight tube sections is predominantly axial, upward with
low cross-flow velocities. Mitsubishi postulated in L5-04GA564, Rev. 2, that turbulent excitation
was the potential cause for wear at tube support plates. The evaluation did not specifically
address the small region of tube wear shown in Figure 2-6 that was observed at tube support
plates 1 through 7.
The correlation between boiling in a small region at the bottom of the tube bundle, based on
Mitsubishi Document calculations, and the observed region of tube wear increasing from tube
support plate 1 levels upward into the U-bend region has apparently not been addressed.
Some mechanism is moving the tubes and causing the tube-to-tube support plate wear in a
small region. Standard steam generator evaluation procedures may not model in sufficient
detail the unusual flow pattern evolving from the tubesheet.
A2-10
6.2.5 Tube Plugging
It has been proposed in Document L5-04GA571, Screening Criteria for Susceptibility to In-
Plane Tube Motion, Revision 4, to plug the tubes exhibiting wear and surrounding tubes that
have a susceptibility for damaging tube motion and freespan wear. The draft AIT inspection
report that was available during this review addressed tube plugging to contain tube damage,
but does not address the specific plugging strategy more recently proposed by Mitsubishi.
Tubes proposed to be plugged in Unit 3 and Unit 3 SGs are selected based on pre-damage
based scoring system. The identified tubes are located in a relatively small contiguous cross-
section area of the tube bundle. These tubes include tubes shown in Figure 2-6 of Document
L5-04GA564, Revision 2 with TWD wear at the tube support plate 1 level. The region with
severe wear in the U-bend tubes originates in the straight sections above the location where
nucleate boiling starts. The implication of this correlation between the hot spot and tube wear
above the spot has not been evaluated.
Plugging of the selected tubes, identified in L5-04GA571, is intended to permit reactor operation
without failures for a time to be defined. Plugging these tubes also eliminates the specific hot
spot at the tube sheet. Locating a cold region above the tube sheet should reduce fluid
temperatures in the region surrounding the previous hotspot, and prevent boiling at the tube
sheet level, which needs to be confirmed by analysis.
6.3 Summary
In summary, the simplified, postulated scenario leading to the damaging tube vibrations of the
U-bend tubes is as follows:
The secondary flow from the wrapper inlet port to the tube bundle does not penetrate the
bundle because the flow resistance of the replacement steam generator bundle with a
smaller pitch-to-diameter ratios higher than in the original steam generator. No evidence
was found that this change was considered in design.
Nucleate boiling occurs at the tube sheet level with low cross flow, a hot spot location, which
is inconsistent with the replacement steam generator system design parameters.
Above the hot spot, undefined flow conditions cause a small group of tubes to vibrate
starting at the tube support plate1 level. Tube wear progressively increases to the upper
tube support plate 7.
The localized high velocity flow with high void fraction causes the U-bend tube bundle to
vibrate violently in a small region above the TWD wear at the tube support plate levels.
More evaluations would be required to substantiate the postulated scenario as the source of the
high void fraction and velocity in a specific U-bend region.
Mitsubishi has selected groups of tubing to be plugged based on damage screening criteria.
The selected tubes include the tubes with TWD wear at the tube support plate 1 level. Plugging
A2-11
these tubes also eliminates the specific hot spot at the tube sheet. Locating a cold region above
the tube sheet should reduce fluid temperatures in regions surrounding the previous hotspot
and prevent boiling at the tube sheet level.
7. OPERATIONAL IMPACTS
The AIT identified an unresolved item requiring further review pertaining to whether the licensee
appropriately reviewed and dispositioned numerous steam generator loose parts alarms during
Unit 3 operation. Similar steam generator loose parts alarms did not occur during Unit 2
operations in Cycle 16, raising the question of whether the Unit 3 alarms were potentially
indicating steam generator tube-to-tube contact during power operations. It was noted from
review of the licensee RCE report that Westinghouse had performed an analysis of the various
alarms for the licensee. Westinghouse concluded that the vibration and loose parts monitoring
system events for both SGs were the result of true metallic impacts and not false indications
from electrical noise or fluctuations in background noise. The alarm events were noted to be
similar to events that occur when SGs shift during reactor coolant system temperature
transients, but it could not be conclusively stated without additional data that the events were
from the same source. The licensee noted that even with additional data, determination of the
source of impacts could be hindered by the location of the sensors. This comment is related to
the fact that accelerometers were mounted on the support skirt for each replacement steam
generator, a remote location with respect to monitoring internal replacement steam generator
conditions.
During this independent review, it was ascertained that the accelerometer skirt location did not
appear to comply with the requirements of the Design Specification SO23-617-01, Specification
for Design and Fabrication of Replacement Steam Generators for Unit 2 and Unit 3, Revision 4.
Specifically, Section 3.9.3.19, Loose Parts Monitoring Provisions, required mounting pads for
sensors to be installed on the external surface of the inlet side of the channel head and on the
lower shell. One pair of mounting pads (one active and one reserve) was required to be located
with a vertical alignment above the tubesheet, and one pair with a vertical alignment below the
tubesheet. Revision 4 of Design Specification SO23-617-01 was approved on July 28, 2010,
which post dates the Unit 2 return to power on April 13, 2010 after replacement steam generator
installation. The circumstances pertaining to relocation of sensors to a lower sensitivity
measurement location, approval of this change, and the continuing conflict with current design
specification requirements were not available for review.
A2-12
8. REFERENCES
Document
Title
Revision
Condition Report
NN 201836127
Root Cause Evaluation: Unit 3 Steam Generator Tube
Leak and Tube-to-Tube Wear Condition
0
Condition Report
NN 201843216
Root Cause Evaluation: [Unit 2] Steam Generator
[Retainer Bar] Tube Wear
0
NRC
Augmented Inspection Team Report
Draft
L5-04GA564
Tube Wear of Unit 3RSG - Technical Evaluation Report
2
ASME Boiler and Pressure Vessel Code,Section III,
Division1, Appendix N, Section N-1330
1992
L5-04GA510
Thermal and Hydraulic Parametric Calculations
5
L5-04GA521
Three Dimensional Thermal and Hydraulic Analysis,
FIT III Code Analysis
3
SO23-617-01
Specification for Design and Fabrication of RSGs for Units
2 & 3
4
L5-04GA504
Evaluation of Tube Vibration
3
L5-04GA428
Design of Anti-Vibration Bar
5
MHI Report
KAS-20040233
SSPC Code Validation and Qualification Report
3
L5-04GA571
Screening Criteria for Susceptibility to In-Plane Tube
Motion
4
KAS-20040233
SSPC Code Validation and Qualification Report
3
L5-04GA411
Design Report of Tube Support Plate and Stay Rod
7
L5-04GA021
Performance Analysis Report
3
UGNR-SON3-
RSG-057
Extension of Tubesheet PWHT Duration
1
UGNR SON3-
RSG-052
Divider Plate Weld Crack (#3A RSG)
19
UGNR-SON3-
RSG-051
Divider Plate Weld Crack (#3B RSG)
16
Attachment 3
NRC INTERNATIONAL TRAVEL TRIP REPORT
Traveler, Office, Division
Office of Research, Divisions of Systems Analysis, Reactor Systems Code Development Branch
Subject:
Low-Frequency Squeeze Film (SF) Damping Tests
Dates of Travel and Countries/Organizations Visited:
March 11th - 14th, 2012, Ontario, Canada/ AECL Chalk River Testing Laboratory
Summary of Trip
From March 11th - 14th, Mr. Ryan Lantz (R-IV), Mr. Carl Thurston (RES), and Dr. Gopinath
Warrier (Contractor - UCLA Adjunct Professor) inspected and observed Low-Frequency
Squeeze Film (SF) Damping Tests at Atomic Energy Canada Limited (AECL) Chalk River
Testing Laboratory in Ontario, Canada. The purpose of the tests was to support the San Onofre
Unit 2 return-to-service steam generator operational assessment (OA) analyses for San Onofre
Unit 2 as related to Unit 3 tube leak and loss of tube integrity due to tube-to-tube wear in the
upper bundle.1 The analytical method used in the Southern California Edison (SCE) operational
assessment uses an empirical damping correlation based on work by Dr. Michel J. Pettigrew.
Dr. Pettigrew has worked as an AECL senior staff member and is currently Chair of Fluid and
Structure Interaction at the Ecole Polytechnique in Montreal.
The licensees methodology for the OA was developed by their replacement steam generator
contractor, Mitsubishi Heavy Industries (MHI), in conjunction with SCE staff and industry experts
including Dr. Pettigrew. During the NRC review of the licensee submittals, the NRC questioned
the MHI application of the SF damping correlation used to compute tube stability margins for
flow induced vibration. Specifically, the NRC review indicated that the correlation was being
non-conservatively extrapolated to lower frequencies beyond published data, and that the
correlation was being non-conservatively applied in regard to structural geometry. The SF
damping correlation was developed for tube support plates where the geometry provides, to a
certain degree, a confined fluid layer around the tube, i.e., for a support plate with drilled or
broached holes. The NRC maintained that this confined layer is not present at flat bar anti-
vibration bar supports and that the correlation should not be applied to anti-vibration bar
intersections.
The SCE-sponsored AECL testing plan2 was designed to provide justification for (1) SF model
use at lower frequencies, in the range of 5-20 Hz, and (2) SF application to flat bar anti-vibration
bar support structures.
AECLs existing straight single-tube test rig (from previous tests in 19883) had been kept in
storage at the lab and was used for all of the tests. The test rig can hold a straight tube
1 San Onofre Nuclear Generating Station - NRC Augmented Inspection Team Report
05000362/2012007 (ML12188A748)
2 AECL EACL Test Plan, Measurement of Steam-Generator Tube Damping Due to Anti-Vibration
Bar Supports, 153-127370-TP-001 Revision D2.
3 B.S. Kim, M.J. Pettigrew and J.H. Tromp, Vibration Damping of Heat Exchanger Tubes In
Liquids: Effects of Support Parameters, Journal of Fluids and Structures, Vol. 2, pp 593-614,
1988.
A3-2
extending to approximately 4.26 m in length between fixed supports, either mounted in air or
immersed in water. The tube-vibration frequency can be varied by selecting the tube
length/diameter and type of tube support mount. The test rig rigidly allows for adjusting the type
and clearance of tube supports. The test rig provides hardware to excite vibrations in the tube
and instruments to measure and record the local damping response.
The purpose of the inspection was to (1) determine if the testing was adequate to support SCEs
use of the squeeze film damping correlation in their OA, and (2) confirm that if the testing met
the requirements of Appendix B (commercial grade dedication).
A. Test Setup
The rig consists of an inner trough attached to an outer trough. The inner trough is slightly
shorter than the outer trough and both are made from 1/4-thick stainless steel plate. The vertical
walls of the inner trough are reinforced by angled buttresses for maximum rigidity. The ends of
the inner trough are closed so that it may be filled with water. All of the fixtures used to mount to
the tube, instrumentation, and the tube supports are mounted from the top edge of the inner
trough and are secured with set screws. These fixtures span the inner trough so that the tube
can be approximately 75 mm below the surface when the rig is filled with water.
As indicated in Figure 1 (inner trough only), the supports are located in the middle of the span.
In order to qualify the test rig and benchmark the 1988 results, the drilled-hole geometry shown
was employed first and then the flat bar anti-vibration bar geometry (Phase 2) was inserted. In
each case, the support was clamped to the mounting fixture so that it could be adjusted to vary
the tube eccentrically within the drilled hole in attempts to replicate selected 1988 tests. The
majority of the tests were performed with the two opposing flat anti-vibration bars oriented
vertically, and perpendicular to the tube axis. This paired configuration of flat bars reasonably
represents a typical anti-vibration bar intersection in a steam generator U-bend.
The mounting jig of the anti-vibration bar fixture allowed for varied lateral positioning of the anti-
vibration bar in relation to the tube. This adjustment allowed for fine sizing of the gap with the
anti-vibration bars including bringing the tube in contact with one side of the anti-vibration bar,
with the contact ranging from slightly touching up to an applied normal-direction contact force as
high as a 3 Newtons.
The test rig is equipped with two electromagnetic exciters (coil exciters) that are used to
produce excite tube vibrations in both the vertical and horizontal directions, respectively, with no
physical contact between exciter and tube. Both coil exciters are mounted on the same
mounting fixture, at a location that is located approximately one quarter of the distance between
the tube support and the tube end support clamps.
Three eddy current proximity probes were used to measure vibrations of the tube. The probes
are capable of use in air or in water. Two of the probes were used to measure the vertical and
lateral positions of the tube relative to the support surface. The third proximity probe was used
to measure the vertical motion of the tube at the excitation location.
Alloy 800 SG tubing with an outside diameter of 0.625 (15.9 mm) and a wall thickness of 0.044
(1.12 mm) was used for these tests. The anti-vibration bar flat bars used were materials
remaining at MHI from the RSG fabrication. The work was performed in accordance with the
AECL Nuclear Laboratories Quality Assurance Program, which is compliant with ISO 9001:2008
requirements.
A3-3
Although the test rig had been kept intact, much of the test setup and instrumentation was
updated from what was used in the 1988 tests in order to take advantage of advancements in
technology since the original tests were performed. The updated equipment included the
excitation devices, the proximity probes, and the data acquisition systems. The NRC inspection
examined the setup, calibration and operator training and also checked the Commercial Grade
Dedication to confirm that the test results met the quality standards required for safety-related
applications. The goals of the inspection team were to confirm that the tests were setup in
accordance with the test plans, confirm adequacy of the test rig to reproduce the 1988 data, and
independent verification of selected test results.
The testing was performed in three phases. Phase 1 tests were used to verify the set up and
procedures, to check instrumentation, and to develop/check the test and analysis procedures.
This was accomplished by using the drilled-hole setup and comparing the data to the 1988
tests. Phase 2 tests were performed for four different tube support configurations (via position
of the clamping fixtures with active lengths of the tube set at 1.50, 2.0, 3.0, and 4.25 m) to
produce expected fundamental frequencies of approximately 35.4, 20.0, 8.6, and 4.4 Hz,
respectively, when the tube was filled with water. All tests were conducted at room
temperature. Trapped air bubbles were removed from the tube prior to testing in water.
Likewise, water droplets were wiped off the tube and supports and dried before tests in air.
Phase 3 tests were not pre-defined during the initial testing and were intended to be follow-on
tests based on findings/difficulties from the Phase 2 tests.
For each tube configuration series, tests were completed under the following conditions:
without support with the trough empty of water to determine the inherent baseline damping
of the configuration,
without support with trough full of water to determine the viscous damping due to the
presence of water,
tube centered in the anti-vibration bar support with the trough full of water with vertical
excitation,
tube centered in the anti-vibration bar support with the trough full of water with lateral
excitation,
small one-sided tube-to-anti-vibration bar gap with the trough empty and full of water, and
tube deflected at anti-vibration bar by amounts preload forces.
Damping results were determined using two different methods. The first method was the
vibration-decay or log decrement method which measures the vibration decay after the
sinusoidal excitation is ended. The log-decrement method fits the vibration amplitude decline
with time to a logarithmically decaying sinusoidal function.
The second method was the half-power spectral method which is based on a standard
vibration response distribution fit to the spectral response peak of the lowest vibration mode
excited by a continuous band-limited random excitation. The continuous sinusoidal and/or
random excitation power inputs by the shaker, used to generate tube vibration, can then be
calculated and used to determine total modal damping.
A3-4
Once the tube and support have been placed in the desired configuration, the tube is excited to
the prescribed resonant vibration and the subsequent damping is measured and recorded via
the data acquisition system. In each case, the coil exciter is used to build up the excitation
amplitude of the fundamental vibration mode to the maximum amplitude allowed using
sinusoidal excitation at the natural frequency. The excitation is then shut off and the vibration is
allowed to decay. The rate of decay in amplitude is then used to determine damping. In order to
provide more realistic conditions, lateral excitation of the tube was used in combination with
vertical for a few select configurations.
B. Verification of Test Setup and Instrumentation
The setup of the test rig was reviewed and found to conform to the SCE specifications.2 All
documents related to setup and testing activities were reviewed. The documents adequately
described the pre-testing, testing, and post-testing activities. Note that the fluid (air or water)
temperature in the test rig was monitored to ensure that all testing was carried out at room
temperature.
AECLs calibration records for the above test equipment were reviewed. The records indicate
that the equipment was calibrated during the period Feb. 11-19, 2013 and that the calibration
was performed as per ASTM standards. The range and accuracy of the test equipment used
was appropriate for the tests being performed. Minor problems were noted with demineralized
water in the trough causing rust on the anti-vibration bar surface, on the excitator, and on the
proximity devices. The rust was photographed and cleaned and considered to have no impact
on the results. AECL then began water change-outs on a more frequent basis and cleaned off
the rust with each water change.
No problems were found with AECL tests processes or equipment setups or implementation of
the test procedures.
C. Verification of Test Results
The raw data (vibration amplitude vs. time data from three proximity probes) measured during
three of the tests (with drilled plate) was analyzed independently by the team. The log-
decrement method was used to calculate the damping ratio. Table 1 shows a comparison of the
results we obtained with those obtained by AECL/SCE and Kim et al. (1988).3 The data
reduction procedure used by AECL/SCE appeared to be correct.
Table 1. Comparison of damping ratios for drilled plate
Damping Ratio (%)
Test No.
AECL/SCE*
Kim et al.
Warrier &
Dhir*
Comments
Comm PH1-01 ExcL
Rep1
0.017 (10.52
Hz)
0.012 (10.6
Hz)
0.019 (10.6
Hz)
Air w/o tube support
Comm PH1-02 ExcL
Rep1
0.76 (8.65
Hz)
0.86 (8.6
Hz)
0.80 (8.6 Hz)
Water w/o tube support
PH1-
T01_Sine_Lrg_1(PF)
1.23 (8.59
Hz)
1.44 (8.6
Hz)
1.25 (8.6 Hz)
Water with tube support
- using log-decrement method
The results of the Phase 1 qualification tests were examined for consistency with the previous
1988 test, where a value of 0.59 was found for L=19mm, H=1.5mm, and eccentricity ratio=0
A3-5
(Table 23). Our preliminary computations indicated SF damping in the range of 0.40 for this
benchmark case. The team considered this result to be reasonably close and adequate.
The raw data from one of the tests using flat plate anti-vibration bar (with only vertical excitation)
was also analyzed, and the log-decrement method was used to calculate the damping ratio. The
natural frequency was also calculated. Table 2 shows the comparison of the damping ratio (and
natural frequency) calculated to those obtained by AECL/SCE. The results in Table 2 also
confirm that the data reduction procedure used by AECL/SCE is reasonable.
Table 2. Comparison of damping ratios for flat plate anti-vibration bar
Test No.
AECL/SCE*
Warrier &
Dhir*
Comments
PH2-1_T12_Sine_Lrg_1
0.764 (8.72 Hz)
0.77 (8.7 Hz)
Water with anti-
vibration bar
tube support
- using log-decrement method
A comparison of the results given in Table 1 (for water w/o tube support) and Table 2 (water with
anti-vibration bar tube support) shows that when flat plate anti-vibration bars are used and the
tube oscillations are in-plane (vertical, i.e., along the length of anti-vibration bar) any additional
damping due to the supports (anti-vibration bar) is negligible.
Based on periodic briefs with Dr. Vijay Dhir (Contractor - UCLA Dean Mechanical & Aerospace
Engineering), the team requested AECL/SCE to take photos of droplet interface of the test water
to tubing material and to the anti-vibration bar surfaces. The static contact angle was measured
to be 55° +/- 5°, for both tube and anti-vibration bar material. This indicates that water was only
partially wetting the solid surfaces.
In conclusion, the anti-vibration bar test results examined by the inspection team showed that
squeeze film damping is negligible when only vertical excitation (in-plane) is present. However,
further Phase 2 testing provided results (note that Phase 2 was only about 25% completed
when we exited the site on 3/14) that showed non-negligible squeeze film damping. The
increased experimental damping observed in the later tests may be due to (1) addition of mixed
lateral motions or (2) where there is a preload added to tube-to-anti-vibration bar contact.
Regardless, based on this testing it is clear that the squeeze film damping correlation3 does not
apply directly to the anti-vibration bar geometry.
Additionally, the inspection team found no apparent problems with the commercial dedication.
The other participates included:
AECL
Dr. Victor P. Janzen
Bruce A. W. Smith (Test Eng)
Nigel J. Fisher
Dr. Paul Feenstra (Test Eng)
Mike Liu
Tom Yackle
Myles Pawlaczyk
Mike Jasurda
A3-6
Figure 1: Sketch of 1988 Two-span Tube Rig inner Trough with Intermediate Support3
Attachment 4
Report to NRC
Submitted by V. K. Dhir
May 31, 2013
The report is organized in order of the questions listed in the work scope.
A. Review and comment on the strength and weakness of existing methodology that is based on
Connors equation. Note any features of the correlation that are most prone to error and/or that may
be limited in applicability. Comment on potential improvements based on current industry data and
trends.
Studies of tube vibrations induced by fluid flowing parallel and across tube bundles have been
reported in the literature since 1970. Most of the reported studies have been experimental in nature
and only a few theoretical efforts have been carried out to understand the mechanisms of
vibrations. Experimental studies have included both single and two phase adiabatic and diabatic
flows over flexible tubes in a tube bundle. During flow along and across tubes, the coupling of flow
induced force with the time varying displacement of tubes causes the tubes to vibrate. Under
normal flow conditions, the amplitude of these vibrations is small and the vibrations do not impose a
threat to the integrity of tubes. However, at certain flow conditions coupling between fluid induced
forces and structural response may be such that a large jump in the amplitude of the vibrations
occurs. Such a condition is referred to as onset of Fluid-Elastic Instability (FEI). Persistence of FEI
can lead to damage to the tubes and loss of integrity. The flow velocity at which FEI occurs is
termed as the critical velocity. In practice, heat exchangers are designed to avoid onset of FEI with
a considerable margin.
In axial single and two phase flows along the tube bundle, the FEI can occur because of fluid flow
inside or outside of the tubes. The dimensionless velocity (), required for onset of FEI in axial flow
is given by Pettigrew and Taylori as
/
(1)
where, , is the axial flow velocity, , is the tube length, , is the hydrodynamic mass of the tube,
, is Youngs modulus, and, , is the moment of inertia of the area of the tube about the tube axis.
Hardly any FEI data leading to large amplitude vibrations under axial flow conditions has been
reported in the literature. Tube vibrations under axial flow conditions can also occur as a result of
nucleate boiling on the surface (small pressure fluctuations on the surface) and as a result of
pressure fluctuations and turbulence in the bulk flow. However, magnitude of vibrations resulting
from these mechanisms is relatively small.
Mechanisms of vibrations of tubes subjected to cross flow have also been reviewed by Pettigrew
and Taylor1 and Blevinsii. The mechanisms identified for flow induced vibrations include coupling of
the flow induced forces with the structural response, turbulence in the flow, and periodic vortex
shedding. In most heat exchanger applications especially under two phase flow conditions, the
latter two mechanisms are of lesser importance. Under certain flow and support conditions, the fluid
forces can lead to onset of FEI. Generally, it is believed that FEI occurs when energy input by fluid
forces acting on tubes exceeds the energy lost by the tubes to the fluid by damping. Connorsiii was
A4-2
perhaps the first one to report data on FEI of a row of tubes held by piano wires and subjected to
cross flow of air. Connors observed that at the onset of FEI, not only the amplitude of vibrations
increases substantially, the tubes also develop a whirling motion and vibrate in oval orbits. Based
on his single phase data and using dimensional analysis, Connors arrived at the criteria for FEI as
(2)
where , is the critical gap velocity through tube array,, is the natural frequency of the tube, ,
is the diameter of the tube, , is the mass of the tube per unit length including the mass of fluid in
the tube and the hydrodynamic mass, , is the damping factor, which depends on the fluid
conditions, and the manner in which tubes are supported, and, , is the density of the fluid. The
constant and exponent were found by Connors while fitting the data to be 9.9 and 0.5
respectively.
Although Connors obtained the correlation from single phase data, subsequent experimental work
by a number of investigators involving single and two phase flows, different types of tube arrays,
and tube supports has shown that general form of equation (2) still holds. However, different
investigators have found that to correlate their data, the magnitude of constant varies significantly
from the value suggested by Connors.
There are a number of reasons for the large scatter that is observed in the data of various
investigators when it is correlated in terms of Connors equation (2).
i.
The natural frequency of a tube depends on the size, configuration, and mass of the tube, the
manner in which the tube is held, its length, and tube material properties. Any ambiguities in
defining tube parameters will be reflected in variations in calculation of the critical velocity from a
correlation such as equation (2).
ii. The critical velocity,, is the gap velocity of the two phase mixture. Aside from gas and liquid
flow rates, the mixture velocity will depend on void fraction and geometrical arrangement of the
tube array. Any uncertainty in void fraction from different correlations will affect the magnitude of
calculated .
iii. The total mass of the tube includes tube mass, mass of fluid in the tube and the hydrodynamic
mass (added mass or the fluid mass that moves with the tube). The tube mass depends on the
size, thickness and density of tube material. The mass of fluid in the tube depends on the
density of the primary coolant and inside diameter of the tube. Void fraction, densities of the two
phases, the pitch to diameter ratio and arrangement of the tube array (e.g., triangular, square,
etc.) will affect the added or hydrodynamic mass.
iv. The degrees of freedom of the tubes have influence on the observed critical velocity for onset of
FEI. Generally, it is found that the threshold for out-of-plane instability (lift direction) is lower than
that for the in-plane instability (drag direction). Janzen et aliv were the first to experimentally
observe U-tube bundle vibrations both in-plane and out-of-plane under air-water two phase flow
conditions. They noted that instability constant in Connors equation was generally higher than
3 and for in-plane instability the flow velocity was about twice that for out-of-plane instability.
Value of for in-plane instability was also found to be twice that for out-of-plane instability when
A4-3
values of damping factor observed in the experiments were used. However when a nominal
value of 1.5% was used for the damping factor, the in-plane value of was only 17% higher
than the out-of-plane value. This is probably a result of the uncertainty in measurement of
damping factor and variability in the damping factor data of various investigators. Violette et al v
investigated tube arrays preferentially flexible in the in-flow direction and subjected to air-water
two phase cross flow. The tubes were either allowed to be flexible in the flow direction or were
allowed to be axi-symetrically flexible. In-plane (drag direction) instability occurred at higher
velocities than that for a tube flexible in all directions. The critical velocity increased with
increase of stiffness or frequency but not in direct proportion to frequency. Although they found a
value of higher than 3 in Connors equation for FEI in the in-plane direction, (some of the data,
however, could be correlated with ), the increase was attributed to a transition at values of
parameter
.
The transition was identified by noting two different trends in the critical velocity data at onset of
FEI when plotted as a function of
.
FEI phenomenon in a tube bundle in which tube flexibility direction is in general different from
either the flow approach direction or transverse to it has been studied by Khalvatti et al.vi Flow
direction was found to affect the onset of instability. Critical velocity decreased when the flow
was normal to the direction in which the tube was most flexible (lift direction).
v. The total damping factor, , includes contributions from internal material damping, tube support
damping including squeeze film damping, viscous damping, and two phase fluid damping.
Internal material damping is due to internal energy dissipation within a material, support
damping or structural damping includes friction and motion of trapped fluid between tube and
support. Viscous damping is due to fluid drag and viscous dissipation. Because of the
increased compliance of a gas-liquid mixture, two phase fluid leads to additional damping. Two
phase fluid damping depends on the density of the surrounding fluid, void fraction of the two
phase mixture and geometry of the tube array including orientation and spacing of tubes. Since
the critical velocity depends approximately on the square root of the damping factor, any
uncertainty in the damping factor would be reflected accordingly in the calculation of critical
velocity at the onset of FEI. This is further discussed below in the context of determination of
total damping factor.
vi. The two phase mixture density depends on the system pressure and temperature and on the
void fraction. Thus depending on the type of correlation that is used in obtaining void fraction for
given liquid and vapor velocity (unless measured), the calculated critical velocities could differ.
This is especially important at very high void fractions where two phase damping decreases
very rapidly to near zero as void fraction reaches unity.
Two parameters that strongly influence the critical velocity for onset of FEI are natural frequency
and damping factor. They are discussed next.
A4-4
Natural Frequency: Natural frequency for lateral vibrations of a simply supported rod or tube is
given asvii,viii
(3)
where is the mode of excitation, , is beam or tube span, , is the Youngs modulus of elasticity of
the rod or tube material, , is the moment of inertia of the cross sectional area of the rod or tube
about the axis of bending, and , is mass per unit length of the tube including fluid filling the tube.
As the natural frequency varies inversely with the square of the length of the tube, an increase in
unsupported span of the tube will lead to reduction in natural frequency and in turn to reduced
velocity for onset of FEI. In the U-bend region of steam generators anti-vibration bars are placed to
limit the amplitude of out-of-plane (lift direction) vibrations. In case there is no gap between tubes
and anti-vibration bars or the amplitude of vibrations is such that tubes hit the anti-vibration bars,
new nodes are created at anti-vibration bars and span length of the tubes decreases. This in turn
leads to an increase in the natural frequency of the tubes.
Damping: Pettigrew and Taylorix,x, and Pettigrew, et al xi from their review of two phase flow induced
vibration studies, have made design recommendations for avoidance of FEI of tubes subjected to
two phase cross flow. Total damping, as described earlier, includes viscous damping,, internal
material and support damping, , and, two phase mixture damping, . Viscous damping is due to
fluid that adheres to the tube and the damping of two phase mixture is in addition to it. As such an
expression for total damping is written as:
(4)
To obtain viscous damping in two phase flow, Pettigrew and Taylor1 used an expression similar to
that used for single phase while replacing single phase kinematic viscosity with the two phase
kinematic viscosity. Two phase kinematic viscosity,, was defined as
(5)
where , is the liquid kinematic viscosity, , is the gas or vapor kinematic viscosity, and is the
void fraction. An expression for viscous damping in terms of two phase kinematic viscosity was
written as
/
(6)
where, , is the fluid density on secondary side, , is the tube diameter, , is the mass of the tube
per unit length, , is the tube frequency, and is the equivalent diameter. The equivalent diameter
ratio,
, for a triangular array was defined as
A4-5
. .
/
(7)
where is the tube pitch. Table 1 gives the value of calculated for a secondary pressure of about
58 bars or 853 psia as a function of . For primary liquid filled tube including the hydrodynamic
mass, the tube mass was calculated as . .
. Tube frequency was
taken to be 34 sec-1.
Viscous Damping Factor
Table 1
,
4
0
47.6
.
0.2
38.5
.
0.4
29.3
.
0.6
19.8
.
0.8
11.1
.
1.0
1.9
.
Structural (internal or material) damping in steam generator tubes is generally much smaller than
the damping introduced by supports such as tube support plates (TSPs) and anti-vibration bars. As
such structural or internal damping may be neglected. For tubes that pass through holes drilled in
plates or broached holes or through egg crate type of supports, there is a gap between the tubes
and supports. A liquid film resides in the gap at not very high qualities. For a vibrating tube liquid
film may be squeezed as the tube moves towards the solid surface. During its sliding motion in TSP,
a tube may experience viscous shear imposed by the liquid film between two solid surfaces. In case
there is no liquid on the shell side, or the tube has a rocking motion, solid to solid contacts can
occur and lead to friction force in addition to viscous shear. Thus the supports can provide
additional mechanisms for dissipation of energy during vibration of tubes and to the damping
experienced by tubes. Generally tubes are not held axi-symmetric in the gap. This uncertainty along
with the complexity of interaction between the squeeze film, viscous shear and sold to solid friction
has limited the validity of the application of theoretical models such as by Mulcahy.xii
A liquid film can also exist between tubes and anti-vibration bars placed in the U-bend region of
steam generators if there is no contact between the tubes and the anti-vibration bars. For out-of-
plane vibrations (lift direction) of the tubes, squeezing out of the liquid film will occur as the tubes
come closer to or impact the anti-vibration bars. During the in-plane motion the tubes will slide over
the liquid film between anti-vibration bars and tubes. Thus in the latter case only viscous shear
damping will be important. Janzen et al 4 have experimentally studied the effect of flat bar U-bend
restraints (FURs) on in-plane and out-of-plane FEI. They have noted that amplitude of vibrations
was limited by the gap between tubes and FURs. No discernible effect of the presence of FURs and
the size of the gap between tubes and FURs on the onset of in-plane or out-of-plane FEI was
observed. This observation is true as long as no physical contact between tubes and FURs occurs.
In an earlier study Weaver and Schneiderxiii studied the effect of flat bar supports on FEI of U-tube
heat exchangers under cross-flow of air. They found that in the presence of little or no gap between
tubes and flat bar supports, higher nodes for out-of-plane vibrations were created. However, when
4 In practical application to nuclear steam generators the contribution of viscous damping is small and
can be neglected.
A4-6
there was a large gap between the tubes and supports, the tubes behaved as if there was no
support until the amplitude of vibrations was such that tubes started to impact the support.
Instability occurred at low flow velocities. The repeated impact could lead to fretting failure of tubes
with time. However, after the tubes started to impact the supports, an increase in flow velocity
triggered higher nodes of instability.
Pettigrew et al11 have correlated available data for tube support damping due to mechanisms of
squeeze film and friction which presumably includes the contribution of viscous shear. The data
were from both laboratory experiments and field tests and the correlation was developed for multi-
span tubes. Their semi-empirical correlation for multi-span tubes accounted for the thickness of the
support plate relative to the tube span and was written as
.
.
.
(8)
The first term in equation (8) accounts for squeeze film damping whereas the second term for
friction between tubes and tube support. In equation (8), , is the number of tube supports, , is
the natural frequency of tubes, , is the fluid density, , is tube diameter, is the tube mass per
unit length, , is the thickness or width of the support and, , is the characteristic (average) span
length. It should be noted that the squeeze film damping is correlated with inverse of frequency and
the data used in developing the correlation covered a frequency range from 30 to 500 Hz. It should
be noted however, in an earlier work, Kim et alxiv correlated the squeeze film damping data as ..
Because of inverse dependence of squeeze film damping on frequency in eq. (8), very high values
of support damping are given by the correlation at low frequencies ( ). As such one must
be very careful in extending the correlation beyond the range of the available data. In the absence
of additional data it may be prudent to limit the upper value of the maximum squeeze film damping
at a frequency of . Also, it should be noted that the correlation is applicable to tubes passing
through eggcrates and circular support plates and currently no justification exists for applying the
results to flat anti-vibration bars.
Damping factor for two phase flow has been reported by a number of investigators. These damping
factors have been deduced from the data for variation of amplitude of vibrations with frequency.
Often a significant scatter in the data is found because of the uncertainty in the evaluation of
damping factor from the vibration amplitude frequency signal, using half-power bandwidth method.
Also, often it is not clear if other components of damping are included or excluded in the reported
data. Based on the data reported in the literature up to 1994, Pettigrew and Taylor1 have proposed
a correlation for two phase flow damping as
(9)
where is an empirical constant that was chosen to have a value of 5. The function was
correlated as
. .
. . .
(10)
A4-7
.
.
.
.
The parameter was included to account for the effect of variation of surface tension with
temperature and was defined as
°
xxx
(11)
Because of the limited steam-water data at different pressures, the definition of parameter is very
tenuous. Recently Mitra, Dhir and Cattonxv obtained FEI data on flexible tube bundles in both air-
water and steam-water mixtures. From their work, they found little systematic combined effect of
change in surface tension with temperature and presence of vapor versus air in the two phase
mixture. Thus at present no definitive data are available to substantiate the form and the value of
exponent in the expression for proposed by Pettigrew and Taylor. It appears reasonable to
assume to be zero or to have a value of unity.
Critical velocity for onset of FEI: It has been suggested by Pettigrew and Taylor9,10 that for design
purposes, Connors equation (2) could be used to determine the critical velocity for onset of FEI by
assuming and. . Pettigrew and Taylor9 also noted that value of was recommended
for tube bundles with /1.47. However, for / less than. , they recommended the
correlation
.
.
(12)
for / of . , eq. (12) gives a value of about . for . Mohany et alxvi have reported instability
data for multi-span U-tubes under two phase flow of Freon. They conclude that for out-of-plane
instability a value of represents the lower bound of data for tubes with /. . We
assume that value of . is a realistic value for out-of-plane instability in RSGs. It should be
noted that in obtaining critical velocity in the U-bend region, only components of velocity normal to
the tube need to be considered.
Table 2 lists the values of and the dimensionless critical velocity for onset of FEI corresponding to
expected thermal hydraulic conditions on the secondary side of RSGs in Units 2 and 3.
A4-8
Table 2
,
0
47.6
3.74
.
0
-
.
0.54
0.2
38.5
4.35
.
.
-
.
2.1
0.4
29.3
5.31
.
.
-
.
3.22
0.6
19.8
7.28
.
.
-
.
3.76
0.8
11.1
12.08
.
.
-
.
4.04
1.0
1.9
61.5
.
0
-
.
0.81
There is a significant uncertainty in the magnitude of squeeze film and anti-vibration bar support
damping. Table 3 shows the dimensionless critical velocity when support damping of 1% and 2%
are assumed.
Table 3
0
.
2.13
.
1.55
0.2
.
3.06
.
2.26
0.4
.
4.04
.
3.65
0.6
.
4.73
.
4.27
0.8
.
5.48
.
4.81
1.0
.
8.38
.
5.95
The calculated dimensionless critical velocity for onset of FEI is plotted in Fig. 1 as a function of
void fraction, with or without support damping. In this figure the expected two phase mixture velocity
through the steam generator (RSG) tube bundle is also plotted as a function of for an assumed
two phase mass flow rate of / / or // and natural frequency of 34 sec.-1
It is noted that for void fractions less than 0.8 in the U-bend region of the bundle (for cross flow), the
critical velocity for onset of instability is higher than the mixture velocity. As such probability of onset
of FEI will be low. However, for void fractions greater than 0.8, the mixture velocity can be
substantially higher than the critical velocity, indicating a strong possibility of occurrence of FEI. In
RSGs for Units 2 and 3, there is a significant area in the upper U-bend region where calculated void
fractions exceed 0.8. When support damping of 1% and 2% is included, the predicted
dimensionless critical velocity for onset of FEI increases for all values of . However, the increase
is substantial for high void fractions (greater than 0.8) because now the support damping dominates
viscous and two phase damping. With support damping of 1%, the two phase mixture velocity can
A4-9
exceed the critical velocity for onset of FEI when the void fraction approaches unity. (Although this
is not the case for the results plotted in Fig. 1 when support damping of 2% is included in the
calculations.) Thus to reduce the propensity of FEI it is essential that very high void fraction regions
( in the upper U-bend region of RSGs should be avoided. Figure 1 can be used to determine
for individual tubes or clusters of tubes the potential for FEI in Units 2 and 3 RSGs. It should also be
noted that lack of contact forces and damping due to anti-vibration bars may be the dominant
reason for existence of FEI and severe tube damage in Unit 3 RSGs. As a result of FEI, initially out-
of-plane vibrations are expected to occur in the lift direction normal to plane of the tubes. However,
this instability in the absence of restraining force from anti-vibration bars can excite in-plane
vibration that can be the cause of tube to tube damage.
B. MHI methods have included assumption of certain number of inactive (non-contact) anti-vibration
bar supports. This lack of engagement in combination with the flowering phenomena is believed to
be the primary cause of wear in Unit 3. Comment on basis of these causal factors.
Critical velocity for onset of FEI is directly proportional to the natural frequency of the tube. Natural
frequency of the tube depends on the manner it is supported (e.g., fixed) at the ends, span length
between intermediate supports and the extent (size of gaps between tubes and supports) of the
intermediate supports. An expression for natural frequency of vibration normal to the axis of a tube
simply supported at the two ends as noted earlier is given as
(13)
0
1
2
3
4
5
6
7
8
9
10
0
0.2
0.4
0.6
0.8
1
Expected Maximum
Mixture Velocity
Zero Support
Damping
With 2% Support
Damping
With 1% Support
Damping
Fig. 1 Critical velocity for onset of FEI as a function of void fraction.
A4-10
Thus we see that natural frequency of the tube will decrease as the unsupported length increases.
Consequently onset of FEI of tubes of longer unsupported span will occur at lower fluid velocity.
Gap between anti-vibration bars and tubes will limit the amplitude of tube vibrations in the lift
direction (out-of-plane). This is expected to be the direction in which FEI will first occur. Existence of
non-contact anti-vibration bars will lead to increased span length for tube vibration and in turn in
reduced natural frequency. Since critical velocity is proportional to frequency, a corresponding
reduction in the fluid gap velocity required for FEI will occur. Consequently in the presence of non-
contact anti-vibration bars, certain number of tubes in the U-bend region where void fractions and
superficial velocities are high will be prone to early FEI. It should be noted that liquid film between
tubes and non-contact anti-vibration bars will provide some damping due to squeezing of the film for
out-of-plane vibrations. When the amplitude of out-of-plane vibrations is such that tubes hit the anti-
vibration bars, tube contact with anti-vibration bars can lead to addition of new nodes and reduced
span length accompanied by increase in natural frequency. For in-plane (drag direction) vibrations,
sliding of tubes directly over anti-vibration bars or over a liquid film will provide additional damping
but it is expected to be much less than that for squeeze film damping in tube supports.
In summary the presence of non-active anti-vibration bars in Units 2 and 3 and repeated impact of
tubes with anti-vibration bars can be a significant cause of tube wear in Units 2 and 3. The
flowering phenomena can occur due to differential thermal stresses but is not expected to be the
main cause of severe damage to the tubes.
C. Provide some assessment of why Unit 2 and Unit 3 RSGs have behaved so differently? Do you
concur with the premise of anti-vibration bars contact forces not being sufficient on Unit 3?
I see two possibilities. One is that anti-vibration bar gap is large for Unit 3 and as a result out of
plane vibration of tubes occurs unrestrained until the magnitudes of vibrations become large. As
discussed in item B, lack of contact leads to reduced natural frequency and in turn, lower fluid
velocity for onset of FEI. Lack of contact force between tubes and anti-vibration bars and smaller
damping due to the presence of anti-vibration bars can be the major cause of out-of-plane and
consequently in-plane large amplitude vibrations in Unit 3. Other factor that can contribute is larger
gap between tubes and TSPs.
The second possible cause, though much less probable, is the maldistribution of secondary flow
resulting from non-uniformities in primary flow. Early onset of FEI in one part of the tube bundle can
lead to tube instability in other parts. If the fluid velocity exceeds the critical velocity, the amplitude
of vibration of tubes in the lift direction (out-of-plane) can substantially increase causing the tubes to
hit noncontact anti-vibration bars. At onset of instability tubes would start to vibrate in the out-of-
plane mode. Large magnitude out-of-plane vibrations in turn can trigger in-plane instability.
Persistence of this condition over a long period of time can lead to thinning of tube wall with
eventual tube failure. Thus FEI is considered to be the major cause of wear in Unit 3. In-plane
flowering of the tubes due to thermal stresses is not considered to be the primary cause of tube
failure. It may have exacerbated the potential for damage due to large amplitude vibrations.
As discussed in section A, contribution of support damping to total damping can be comparable to
that from two phase mixture. Support damping includes squeeze film damping, viscous shear
damping and solid to solid friction damping. Presence of any or all of these contributors and their
interactions can affect the magnitude of support damping. As such a large scatter is seen in the
data that have been correlated by Pettigrew et al11. However, one should be careful in using the
correlation to support geometries and configurations that prevail in a given situation. For example, it
A4-11
is expected that magnitude of damping from anti-vibration bars will be less than that due to TSPs.
Nevertheless, lack of anti-vibration bar support (no contact) force and little support damping and
larger gaps between TSP and tubes in RSG 3 could have caused it to behave differently than
RSG 2.
D. Review MHI root cause report and provide an assessment of the report findings and conclusions
focusing on anti-vibration bar contact forces (damping impacts associated with Connors equation)
due to manufacturing differences and modeling errors.
As has been discussed earlier in connection with Item A, correlations for squeeze film damping
have been developed from data from tubes passing through drilled and broached holes and egg
crates. The geometrical configuration of tubes passing through holes is very different than that of a
tube placed adjacent to a flat surface of anti-vibration bars. Because of the geometrical and
orientation differences, the magnitudes of squeeze film and viscous shear damping in anti-vibration
bars are expected to be different. For out-of-plane vibration of tubes in RSGs, squeeze film
damping will occur but its magnitude has not yet been quantified. As the tubes vibration amplitude
equals the gap between tubes and anti-vibration bars, the tubes will start to impact the anti-vibration
bars. This in turn will cause damage to the tubes and lead to inactive anti-vibration bars becoming
active and creating additional vibrational nodes. For in-plane movement the damping due only to
viscous shear caused by sliding of tubes against a liquid film between tubes and anti-vibration bars
will take place. In this case, little contribution will come from squeeze film damping. One could
theoretically model viscous shear contribution which amongst other variables will depend on the
width of anti-vibration bars. However, any such model predictions must be validated under
prototypical conditions.
E. Impact of recent testing by (CANDU Services 147-02120-505452-371-9001) AECL on the damping
due to anti-vibration bars and return to service of SONGS RSGs.
In response to the questions that we (NRC) asked in connection with SONGS application for return
to service of Unit 2, SONGS requested AECL to carryout tests to document the vibrational damping
introduced by anti-vibration bars. Two key questions we had raised were:
1. What is the rationale for applying squeeze film damping correlations developed based on the
data for circular supports to flat anti-vibration bars?
2. In the absence of any supporting data, what is the rationale for extrapolating squeeze film
damping correlations for circular supports to tube frequencies less than 30?
Tests for damping factors for circular supports and anti-vibration bars have been performed by
AECL. In these tests prototypical single tubes of lengths varying from 4.2 meters to 1.5 meters have
been used. The tubes have been placed in air or a pool of water at about room temperature and are
externally excited. In the tests both in-plane and out-of-plane vibrational modes have been
investigated. The width of gaps between tubes and supports has been varied parametrically.
Damping factors with point values have been reported and the key observations from review of the
data are:
1. The damping factors observed for squeeze film damping in circular supports are found to be a
function of amplitude of vibrations and generally increase with increase in amplitude. For
amplitude to gap width ratio of about 60%, the observed damping factors for tube frequencies of
31.9 and 8.7 and gap radial widths of 0.75 and 0.38mm are less than 1% and are also less than
A4-12
those obtained from the correlations reported in the literature and used by SONGS. Although
limited in scope, damping factors for a given gap width do not scale inversely with the frequency.
One should note that these damping factors have been obtained with water at an atmospheric
pressure and at about room temperature. During RSG operation pressure and temperature will
be much higher. With increase in temperature liquid viscosity will decrease and as a result
squeeze film damping factors may decrease further from those measured at room temperature.
2. For in-plane vibrations in the presence of anti-vibration bars, the damping is found to increase
marginally. There is significant scatter in the data. The data show an increase in damping with
frequency and decrease with gap size. For the highest frequency (31.9 ) and smallest gap
(0.0025 mm), the presence of anti-vibration bars in water leads to a maximum increase in
damping factor due to sliding of the tube parallel to anti-vibration bars of less than 0.1 - 0.15%.
3. In experiments with out-of-plane vibrations in which vibrating tube pushes against the anti-
vibration bars and in turn acts to squeeze out the intervening liquid film, damping factors have
been observed to be higher than those in which tubes were only excited in-plane.
Measurements were made for gap widths of 0.125mm, 0.025mm and 0.0025mm. For 0.125mm,
the highest value of 0.45% for increase in damping factor was noted for RMS displacement to
gap ratio of 1.4 and tube frequency of 4.5 . Tube with a natural frequency of 8.7 yielded an
increase of about 0.18% in damping ratio for RMS displacement/gap ratio of 2. For a gap of
0.0025mm, an increase in damping ratio of about 3% was observed for a tube with 4.54
natural frequency and RMS displacement of 0.16mm. However, the data for the three gaps
studied do not show a consistent trend with tube frequency. These data show that when out-of-
plane motion exists, damping due to anti-vibration bars is higher than that for purely in-plane
vibrations. However, in most cases, additional damping due to anti-vibration bars is less than 1%.
4. Damping factors, friction force and friction coefficient data have been reported when tubes are
pre-loaded. Damping factors have been found to increase with pre-load as friction force between
tubes and anti-vibration bars increases. The damping factors with friction have been found to
increase as tube vibration frequency is decreased but decrease as the amplitude of vibrations is
increased.
5. In terms of consequences of the above described experimental effort on the discussion in the
earlier parts of the report, it is concluded that for in-plane tube vibrations, anti-vibration bars
contribute little to additional damping. However, for out-of-plane vibrations, the additional
damping (squeeze film) due to anti-vibration bars is higher. Although data show significant
scatter, in most cases in the absence of any physical contact between tube and anti-vibration
bars and any preload, the additional damping (squeeze film) due to anti-vibration bars is less
than 1%. Furthermore this does not account for the reduction of viscosity of water with increase
in temperature. Thus in Figure 1, the predicted critical velocity curve for 1% support damping
seems to be more appropriate. Accordingly, with the information we have at this point, a certain
number of tubes in the upper portion of the U-tube bundle could experience FEI. However, the
exact number of tubes experiencing FEI would have to be determined by knowing the predicted
flow velocities, frequencies and void fraction in each region of interest at 70% power.
A4-13
References
i Pettigrew, M. J. and Taylor, C.E., Two-Phase Flow-Induced Vibration: An Overview, ASME Journal of Pressure
Vessel Technology, Vol. 116, pp 233-253, 1994.
ii Blevins, R.J., Flow Induced Vibration, Second Edition, Van Nostrand Reinhold, Co., New York, 1980.
iii Connors, H.J., Fluid-elastic Vibration of Tube Arrays by Cross Flow, Flow Induced Vibrations in Heat
Exchangers, ASME Winter Annual Meeting, 1970.
iv Janzen, V.P., Kagberg, E.G., Pettigrew, M. J., and Taylor, C.E., Fluidelastic Instability and Work-Rate
Measurements of Steam-Generator U-Tubes in Air-Water Cross-Flow, Journal of Pressure Vessel Technology,
Vol. 127, pp 84-81, 2005.
v Violette, R., Pettigrew, M. J., and Mureithi, N. W., Fluidelastic Instability of an Array of Tubes Preferentially
Flexible in the Flow Direction Subjected to Two Phase Cross Flow, Journal of Pressure Vessel Technology, Vol.
128, pp 148-159, 2006.
vi Khalvatti, A., Mureithi, N.W., and Pettigrew, M. J., Effect of Preferential Flexibility Direction on Fluidelastic
Instability of a Rotated Triangular Tube Bundle, Journal of Pressure Vessel Technology, Vol. 132, Issue 4,
pp 041309 (14 pages), 2010.
vii Dukkipati, R.V., Mechanical Vibrations, Alpha Science International, Limited, Oxford, UK, 2010.
viii Case, J., Chilver, L. and Ross, Carl, T.F., Strength of Materials and Structures, Arnold Publishers, London, UK,
1999.
ix Pettigrew, M.J., and Taylor, C.E., Vibration Analysis of Shell-and-tube Heat Exchangers: An Overview - Part 2:
Flow, Damping, Fluidelastic Instability, Journal of Fluids and Structures, Vol. 18, pp 469-183, 2003.
x Pettigrew, M.J., and Taylor, C.E., Vibration Analysis of Steam Generators and Heat Exchangers: An Overview,
Proceedings of IMECE 2002, New Orleans, Louisiana.
xi Pettigrew, M. J., Rogers, R. J., and Axisa, F., Damping of Heat Exchanger Tubes in Liquids: Review and Design
Guidelines, Journal of Pressure Technology, Vol. 133, Issue 1, 014002 (11 pages), 2011.
xii Mulcahy, T.M., Fluid Forces on Rods Vibrating in Finite Length Annular Regions, Journal of Applied Mechanics,
Vol. 47, pp 234-240, 1980.
xiii Weaver, D.S. and Schneider, W., The Effect of Flat Bar Supports on the Crossflow Induced Response of Heat
Exchanger U-Tubes, Journal of Engineering and Power, Vol. 105, pp 775-781, 1983.
xiv Kim, B.S., Pettigrew, M.J., and Tromp, J.H., Vibration Damping of Heat Exchanger Tubes in Liquids: Effects of
Support Parameters, Journal of Fluids and Structures, Vol. 2, pp 593-613, 1988.
xv Mitra, D., Dhir, V.K., and Catton, I., Fluid-elastic instability in tube arrays subject to air-water and steam-water
cross-flow, Journal of Fluids and Structures, Vol. 25, pp 1213-1235, 2009.
xvi Mohany, A., Janzen, V.P., Feinstra, P., and King, S., Experimental and Numerical Characterization of Flow-
Induced Vibration of Multispan U-tubes, Journal of Pressure Vessel Technology, Vol. 134, pp 013301-1 to 9,
Feb, 2012.
Attachment 5
PRELIMINARY SIGNIFICANCE DETERMINATION
LOSS OF STEAM GENERATOR TUBE INTEGRITY
Assumptions
1. Unit 3 operated for 11.3 effective full power months with newly-installed steam generators prior
to the tube leak. Using a t/2 assumption, the exposure period is assumed to be 5.65 months, or
172 days.
2. The risk of the condition consists of two elements:
a) The increased frequency for a steam generator tube rupture, given the degraded state of the
tubes.
b) A main steam line break could have occurred during the exposure period, resulting in one or
more tubes rupturing, resulting in a compound accident.
3. All steam line breaks that hypothetically could have occurred during the exposure period are
assumed to result in a steam generator tube rupture, given no operator actions. Operators can
take immediate actions according to their procedures that would mitigate the differential
pressure across the tubes. These actions are primarily to prevent a re-pressurization of the RCS
by cooling down using the intact steam generator and eventually shutting off any operating
charging pumps. The probability that these actions will fail to preclude a tube rupture was
assessed using the SPAR-H methodology, as follows:
Diagnosis (0.01)
Action (0.001)
Available
Time
Nominal (1.0)
Time available = time required (10)
Complexity
Moderate (2)
Nominal (1.0)
Stress
High (2.0)
High (2.0)
Procedures
Nominal (1.0)
Nominal
(1.0)
0.04
0.02
Total HEP1
0.06
Note 1: human error probability
All other performance shaping factors are assumed to be nominal.
4. It is assumed that a steam generator tube rupture that results in core damage will always result
in a large early release because all of the radiation barriers, including containment, are
bypassed. This is a bounding assumption because there are some actions the plant can take
that would limit the release to levels that would not meet the definition of a large, early release.
Large early release frequency is defined as the frequency of those accidents leading to
significant, unmitigated releases from containment in a time frame prior to effective
evacuation of the close-in population such that there is the potential for early health
effects. Such accidents generally include unscrubbed releases associated with early
containment failure shortly after vessel breach, containment bypass events, and loss of
containment isolation.
A5-2
5. The degraded condition of the steam generator tubes was first manifested by a small leak that
had little effect on core damage or large early release. However, there was a possibility that a
tube rupture could have occurred as the first manifestation of a problem. For this analysis, it is
assumed that the frequency of a steam generator tube rupture was doubled from the baseline
value for the 172-day exposure period. The baseline value is 2.07E-3/year (1 in 483 years) and
it was doubled to 4.1E-3/year (1 in 241 years).
6. The San Onofre SPAR model, Revision 8.22, was used for this analysis, assuming average test
and maintenance, and a truncation limit of 1.0E-11.
Analysis
Steam Line Break/Steam Generator Tube Rupture
During the 172-day exposure period, if a steam line break had occurred, it is assumed that a
steam generator tube rupture would have occurred concurrently, given no operator action.
According to Assumption # 3 above, the probability that operator actions would fail to preclude a
tube rupture is 0.06. If the steam line break does not cause a tube rupture, it is a baseline event
having no significance to this finding.
It is most likely that only one tube would have ruptured, given that two of the three degraded
tubes were marginally close to passing the test. This places the thermo-hydraulic
considerations consistent with a steam generator tube rupture event.
The SPAR model event tree for main steam line breaks contains 8 sequences. Sequences 3, 4,
6, and 7 are core damage sequences. Sequences 1, 2, and 5 are OK sequences. Sequence 8
is a transfer to an anticipated transient without scram (ATWS) event tree.
The core damage sequences and the ATWS transfer were quantified with a result of 1.61E-6/yr.
The delta-CDF for these sequences is zero (baseline) for this finding because core damage
would have occurred even without the performance deficiency, but because of the possibility of
a concurrent steam generator tube rupture, the delta-LERF must be considered. The baseline
LERF for a steam line break is zero for this plants containment structure, so any core damage
resulting from a concurrent steam generator tube rupture constitutes a delta-LERF. According
to the HEP developed in Assumption #3, the delta-LERF for the core damage sequences is
1.61E-6 (0.06) = 9.65E-8/yr. For a 172-day exposure, this results in an ICLERP of 4.55E-8.
The OK sequences of the main steam line break sequences add to both the delta-CDF and
the delta-LERF because they can result in a tube rupture with a probability of core damage from
that situation independently. Therefore, main steam line break sequences 1, 2, and 5 were
transferred to the steam generator tube rupture event tree. These sequences were quantified
as 0.3825, 0.1275, and 0.3675 respectively, multiplied by the steam line break initiating event
frequency of 7.7E-3.
The OK sequences of the main steam line break event tree all include a success of either
main feedwater or auxiliary feedwater. Therefore, the core damage sequences of the steam
generator tube rupture event tree that involve a failure of feedwater were not quantified for this
portion of the analysis. This left steam generator tube rupture core damage sequences #s 3, 6,
7, 14, 15, 17, 18, and 20 to be quantified, with a result of a CCDP of 2.0E-4.
A5-3
The same operator recovery of an HEP of 0.06 applies in this case.
The delta-CDF and delta-LERF resulting from a main steam line break that would normally have
not resulted in core damage is therefore:
(0.3855 + 0.125 + 0.3675) (7.7E-3) (0.06) (2.0E-4) = 8.10E-8/yr. For a 172-day exposure, this
results in an ICCDP and ICLERP of 3.82E-8.
The following table presents the total risk resulting from main steam line breaks:
Steam line breaks that directly result
in core damage
0
4.55E-8
Steam line breaks that would
normally not result in core damage,
but can cause a steam generator
tube rupture
3.82E-8
3.82E-8
TOTAL
3.82E-8
8.37E-8
Independent Steam Generator Tube Rupture
According to Assumption #5, it is assumed that the frequency of a steam generator tube rupture
was twice the normal rate during the 172-day exposure period. This is not based on an
analytical analysis, but represents a reasonable estimate.
The SPAR model nominal case delta-CDF for a steam generator tube rupture is 4.2E-7/yr.
Therefore, the delta-CDF and delta-LERF resulting from independent steam generator tube
ruptures would also be 4.2E-7. For a 172-day exposure period, the Delta-CDF and Delta-LERF
was 2.0E-7.
Total Internal Risk
Delta-CDF
Delta-LERF
Breaks
3.8E-8
8.4E-8
Independent Steam
Generator Tube
Ruptures
2.0E-7
2.0E-7
Total Internal Risk
2.4E-7
2.8E-7
Risk Significance
Green
White
The finding was green for the increase in core damage, but white for the increase in large early
release.
External Events: Core damage frequency for high winds, floods, was not quantified in
licensees IPEEE because these events met the NRC screening criteria.
A5-4
Seismic and fire events were possible contributors to CDF and LERF. However, the licensees
IPEEE stated that the median capacity of the steam generators was 8g. This was substantial
when compared to other plant components. The design bases earthquake for SONGS was
0.67g.
The licensees risk assessment stated, in part:
The potential for a seismically induced steam generator tube rupture was evaluated as part of
the overall safety significance of the Unit 3 event. The deterministic analysis concluded that
the most limiting degraded tube on Unit 3 would have been able to withstand a design basis
earthquake based on in-situ test results. Therefore, the condition of the tubes was not
considered an additional risk contributor in the PRA. More information on the seismic tube
analysis is provided in Section 5.2.1 of San Onofre's Return to Service Report (which is
enclosure 2 to San Onofre's Confirmatory Action Letter Response Letter).
The best estimate for the change to CDF and LERF from external events was 0.
External Events Sensitivity Studies: To evaluate fire and seismic events to the extent
practicable, the analysts performed two risk evaluations considering different scenarios. The
first considered the potential for a seismic induced tube rupture caused by the weakened tubes.
Since fire scenarios would not directly cause a weakened tube to rupture, fire was not a
contributor in this evaluation. The second evaluation increased the post core damage risk
associated with the effect of a hot gas layer on the weakened tubes. Seismic and fire scenarios
were considered here. Both evaluations required the use of significant judgment when specific
statistics were lacking. Both evaluations were very conservative and bounding.
Seismic Induced Tube Rupture: The analyst considered an increase in core damage and large
early release from a seismic induced tube rupture. The rupture is caused by vibrating the steam
generators and rupturing a weakened portion of a tube.
Using the SPAR model, the CCDP of a steam generator tube rupture combined with an
unrecoverable loss of offsite power (assumed in a large seismic event) was 1.64E-3.
The analyst qualitatively considered that the fragility of the degraded tubes would be at the
approximate g-level earthquake that causes a 25% chance of a small break loss of coolant
accident. This figure was 0.7 g, according to NUREG/CR-4840, Figure 3-6. The frequency of an
earthquake of this magnitude or greater at San Onofre is 1.0E-4/yr.
Therefore, the delta-CDF and delta-LERF from a seismic event is approximately 1.6E-3 (1.0E-
4/yr) = 1.6E-7/yr. For a 172-day exposure, the delta-CDF and delta-LERF was 7.7E-8.
Post-Core-Damage Hot Gas Layer: The SONGS Individual Plant Examination of External
Events (IPEEE), dated December 15, 1995, reported the core damage frequency (CDF) for fire
and seismic external events at SONGS was 3.3E-5 per year. The seismic contribution was
1.7E-5/yr. The seismic risk included accelerations beyond the SONGS design basis earthquake
of 0.67 g peak ground acceleration.
The SONGS containment and pressure boundary response was reported in the IPEEE
(Table 3.7-1). The SGTR can occur post-core-damage when the hot gas layer migrates to the
A5-5
steam generator tubes. The high reactor coolant system pressure coincident with the heat
weakened tubes results in the tube failure.
The analyst qualitatively assumed that the degraded steam generator tubes would make the
tube failure twice as likely as assumed in the licensees IPEEE (from 2% to 4%).
The increase in LERF would be the total external event core damage frequency multiplied by
the change in the probability of SGTR duration fraction, and the fraction of exposure time.
Delta-LERF = 3.3E-5 X (.04 -.02) X 0.47 = 3.1E-7/yr
Total Risk, Best Estimate
Delta-CDF
Delta-LERF
Internal
2.4E-7
2.8E-7
External
0
0
Total
2.4E-7
2.8E-7
Classification
Green
White
Sensitivity for External Events - Vibration Induced Tube Rupture
Delta-CDF
Delta-LERF
Internal
2.4E-7
2.8E-7
External
7.7E-8
7.7E-8
Total
3.1E-7
3.6E-7
Classification
Green
White
Sensitivity for External Events - Hot Gas Layer Failures
Delta-CDF
Delta-LERF
Internal
2.4E-7
2.8E-7
External
3.1E-7
3.1E-7
Total
5.5E-7
5.9E-7
Classification
Green
White
Licensee Analysis
The licensee analysis was presented in PRA Report PRA-12-007, and used a different
analytical technique using studies performed by a contractor used to predict a tube failure
probability as a function of time, assuming linear degradation. The result was a Delta-LERF of
1.4E-7. This is also a White significance.