ML021260141: Difference between revisions

From kanterella
Jump to navigation Jump to search
(StriderTol Bot insert)
 
(StriderTol Bot change)
 
Line 1: Line 1:
#REDIRECT [[IR 05000346/2002003]]
{{Adams
| number = ML021260141
| issue date = 05/03/2002
| title = IR 05000346/2002-003, Davis-Besse Nuclear Power Station. Inspection Conducted on 03/12-04/05/2002, Augumented Inspection Team
| author name = Dyer J
| author affiliation = NRC/RGN-III/DNMS/FCB
| addressee name = Bergendahl H
| addressee affiliation = FirstEnergy Nuclear Operating Co
| docket = 05000346
| license number = NPF-003
| contact person =
| document report number = IR-02-003
| document type = Inspection Report, Letter
| page count = 41
}}
See also: [[see also::IR 05000346/2002003]]
 
=Text=
{{#Wiki_filter:May 3, 2002
Mr. Howard Bergendahl
Vice President - Nuclear
FirstEnergy Nuclear Operating Company
Davis-Besse Nuclear Power Station
5501 North State Route 2
Oak Harbor, OH  43449-9760
SUBJECT:
DAVIS-BESSE NUCLEAR POWER STATION
NRC AUGMENTED INSPECTION TEAM - DEGRADATION OF THE
REACTOR PRESSURE VESSEL HEAD - REPORT NO. 50-346/02-03(DRS)
Dear Mr. Bergendahl:
Your staff provided information to the NRC between March 6 and 10, 2002, concerning the
identification of a large cavity in the reactor vessel head adjacent to a control rod drive nozzle.
On March 13, 2002, the NRC issued a Confirmatory Action Letter outlining specific actions your
staff are expected to take in response to this event.  One of those actions is obtaining NRC
approval prior to restart of the Davis-Besse plant.
On March 12, 2002, the NRC dispatched an Augmented Inspection Team (AIT) to the Davis-
Besse site in accordance with NRC Management Directive 8.3, NRC Incident Investigation
Program.  The AIT was chartered to determine the facts and circumstances related to the
significant degradation of the reactor vessel head pressure boundary material.  The AIT
developed a sequence of events, interviewed plant personnel, collected and analyzed factual
information relevant to the degraded condition and conducted visual inspections of the reactor
vessel head.  The enclosed report provides the AIT findings which were summarized for you
and your staff during a public exit meeting on April 5, 2002.
The cavity in the reactor vessel head was discovered during maintenance activities for
problems found during inspections conducted pursuant to NRC Bulletin 2001-01,
Circumferential Cracking of Reactor Pressure Vessel Head Penetration Nozzles.  The
degraded area covers approximately 30 square inches where the thick low-alloy structural steel
was corroded away, leaving only the thin stainless steel cladding layer as a pressure boundary
for the reactor coolant system.  This represents a loss of the reactor vessels pressure retaining
design function, since the cladding was not considered as pressure boundary material in the
structural design of the reactor pressure vessel.  While the cladding did provide a pressure
retaining capability during reactor operations, the identified degradation represents an
unacceptable reduction in the margin of safety of one of the three principal fission product
barriers at the Davis-Besse Nuclear Power Station.
The AIT concluded that the cavity was caused by boric acid corrosion from leaks through the
control rod drive nozzles in the reactor vessel.  These leaks were caused by primary water
stress corrosion cracking of the nozzle material leading to a through-wall crack and corrosion of
low alloy steel that went undetected for an extended period of time.  The boric acid corrosion
 
H. Bergendahl
-2-
control program at the site included both cleaning and inspection requirements, but was not
effectively implemented to detect the leakage and prevent the significant corrosion of the
reactor vessel head over a period of years.  Similarly on several occasions, maintenance and
corrective action activities failed to detect and address the indications in the containment that
the significant corrosion of the reactor vessel head was occurring.  The NRC views these as
missed opportunities to identify and correct this significant degradation to the reactor pressure
vessel head.
The AIT did not address the verification of compliance with NRC rules and regulations, provide
recommendations for enforcement actions, or assess the risk significance of this issue.  A
followup special inspection effort will be scheduled in the near future to pursue these aspects of
the regulatory process.
In accordance with 10 CFR 2.790 of the NRCs "Rules of Practice," a copy of this letter and
its enclosures will be available electronically for public inspection in the NRC Public
Document Room or from the Publicly Available Records (PARS) component of NRCs
document system (ADAMS).  ADAMS is accessible from the NRC Web site at
http://www.nrc.gov/NRC/ADAMS/index.html (the Public Electronic Reading Room).
Sincerely,
/RA by J. L. Caldwell for/
J. E. Dyer
Regional Administrator
Enclosure:
NRC Augmented Inspection Report
  No. 50-346/02-03(DRS)
cc w/encl:
B. Saunders, President - FENOC
Plant Manager
Manager - Regulatory Affairs
M. OReilly, FirstEnergy
Ohio State Liaison Officer
R. Owen, Ohio Department of Health
Public Utilities Commission of Ohio
 
H. Bergendahl
-2-
control program at the site included both cleaning and inspection requirements, but was not
effectively implemented to detect the leakage and prevent the significant corrosion of the
reactor vessel head over a period of years.  Similarly on several occasions, maintenance and
corrective action activities failed to detect and address the indications in the containment that
the significant corrosion of the reactor vessel head was occurring.  The NRC views these as
missed opportunities to identify and correct this significant degradation to the reactor pressure
vessel head.
The AIT did not address the verification of compliance with NRC rules and regulations, provide
recommendations for enforcement actions or assess the risk significance of this issue.  A
followup special inspection effort will be scheduled in the near future to pursue these aspects of
the regulatory process.
In accordance with 10 CFR 2.790 of the NRCs "Rules of Practice," a copy of this letter and
its enclosures will be available electronically for public inspection in the NRC Public
Document Room or from the Publicly Available Records (PARS) component of NRCs
document system (ADAMS).  ADAMS is accessible from the NRC Web site at
http://www.nrc.gov/NRC/ADAMS/index.html (the Public Electronic Reading Room).
Sincerely,
J. E. Dyer
Regional Administrator
Enclosure:
NRC Augmented Inspection Report
  No. 50-346/02-03(DRS)
cc w/encl:
B. Saunders, President - FENOC
Plant Manager
Manager - Regulatory Affairs
M. OReilly, FirstEnergy
Ohio State Liaison Officer
R. Owen, Ohio Department of Health
Public Utilities Commission of Ohio
DOCUMENT NAME:  G:DRS\\ML021260141.wpd
To receive a copy of this document, indicate in the box:  "C" = Copy without attachment/enclosure "E" = Copy with attachment/enclosure "N" = No copy
OFFICE
RIII
RIII
RIII
RIII
RIII
NAME
JGrobe for
RGardner:sd
CLipa   
JStrasma       
JGrobe         
JCaldwell
for JDyer
DATE
05/03/02
05/03/02
05/03/02
05/03/02
05/
OFFICIAL RECORD COPY
 
H. Bergendahl
-3-
ADAMS Distribution:
Chairman Meserve
Commissioner Dicus
Commissioner Diaz
Commissioner McGaffigan
Commissioner Merrifield
W. Travers, EDO
W. Kane, OEDO
H. Nieh, OEDO
OCIO
ACRS
S. Collins, NRR
B. Sheron, NRR
J. Zwolinski, NRR
J. Strosnider, NRR
G. Holahan, NRR
B. Bateman, NRR
S. Bajwa, NRR
S. Long, NRR
OPA
OCA
H. Miller, RI
L. Reyes, RII
E. Merschoff, RIV
AJM
DFT
SPS1
RidsNrrDipmIipb
HBC
C. Ariano (hard copy)
DRPIII
DRSIII
PLB1
JRK1
 
U.S. NUCLEAR REGULATORY COMMISSION
REGION III
Docket No:
50-346
License No:
NPF-3
Report No:
50-346/02-03
Licensee:
FirstEnergy Nuclear Operating Company
Facility:
Davis-Besse Nuclear Power Station
Location:
5501 North State Route 2
Oak Harbor, OH 43449
Dates:
March 12 - April 5, 2002
Team Leader:
R. Gardner, Engineering Branch Chief, DRS
Inspectors:
J. Davis, Sr. Materials Engineer, RES
M. Holmberg, Sr. Reactor Inspector, DRS
J. Gavula, Sr.  Reactor Inspector, DRS
D. Simpkins, Resident Inspector, Davis-Besse, DRP
Approved by:
John A. Grobe, Director
Division of Reactor Safety
 
ii
TABLE OF CONTENTS
PAGE
SUMMARY OF FINDINGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
1.0
BACKGROUND AND EVENT OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1
Description of Reactor Vessel Head and Penetration Nozzles . . . . . . . . . . . . . . 1
1.2
Sequence of Events:  Discovery of Reactor Vessel Head Degradation . . . . . . . 2
2.0
CHARACTERIZATION OF NOZZLE CRACKING AND REACTOR VESSEL HEAD
WASTAGE AREAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.1
CRDM Nozzle Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2
Reactor Vessel Head Wastage Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
3.0
PROBABLE CAUSE OF NOZZLE CRACKING AND HEAD WASTAGE . . . . . . . . . . . . 4
3.1
Probable Cause for Nozzle Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3.1.1
Factors Affecting Primary Water Stress Corrosion Cracking of Nozzles . 4
3.1.2
CRDM Nozzle Materials and Contributing Factors . . . . . . . . . . . . . . . . . 5
3.2
Probable Cause for Vessel Head Wastage Cavities . . . . . . . . . . . . . . . . . . . . . 6
3.2.1
Boric Acid Corrosion Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.2.1.1
Boric Acid Corrosion Processes . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.2.1.2
Industry Accepted Boric Acid Corrosion Rates . . . . . . . . . . . . . . 7
3.2.2
Licensee Preliminary Identified Cause . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.0
HISTORY OF VESSEL HEAD INSPECTIONS AND MATERIAL CONDITION . . . . . . . 8
4.1
Background CRDM Flange Leakage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.2
History of Flange Leakage and Reactor Head Inspections
. . . . . . . . . . . . . . . . 9
5.0
OPPORTUNITIES FOR EARLY DETECTION OF HEAD DEGRADATION . . . . . . . . . 12
5.1
Boric Acid Corrosion Control Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
5.2
Reactor Coolant System Leakage Detection . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.3
Containment Air Coolers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.4
Radiation Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.5
Causal Factors Influencing Head Degradation Detection . . . . . . . . . . . . . . . . . 18
5.5.1
Decision to Delay Modification to Service Structure . . . . . . . . . . . . . . . 18
5.5.2
Decision to Delay Repair of CRDM Flange on Nozzle 31 in 11RFO . . . 19
6.0
CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
ATTACHMENTS
Attachment A -
USNRC Memorandum from J. E. Dyer to R. N. Gardner, dated March 12,
2002:  Augmented Inspection Team Charter - Davis-Besse Reactor Vessel
Head Material Loss
Attachment B -
NRC Briefing Slides for the Public AIT Exit Meeting Conducted April 5, 2002
Attachment C -
FirstEnergy Intra-Company Memorandum from S. A. Loehlein to H. W.
Bergendahl, dated March 22, 2002
 
iii
SUMMARY OF FINDINGS
IR 05000346-02-03, on 03/12-04/05/2002, FirstEnergy Nuclear Operating Company,
Davis-Besse Nuclear Power Station.  Augmented Inspection Team.
This report covers a 3-week inspection by an NRC Augmented Inspection Team for the
substantial loss of material from the reactor pressure vessel head.
*
On March 5 and 6, 2002, workers at Davis-Besse were repairing control rod drive
penetration Nozzle 3, following the identification of cracks detected through inspections
performed pursuant to NRC Bulletin 2002-01, Circumferential Cracking of Reactor
Pressure Vessel Head Penetration Nozzles.  The workers discovered a large cavity, a
significant loss of metal adjacent to the control rod drive nozzle in the reactor vessel
head, that apparently resulted from boric acid corrosion of the reactor vessel head due
to leakage from the cracks in Nozzle 3.
*
The cracks in the control rod drive nozzles were apparently due to primary water stress
corrosion cracking of the Alloy 600 nozzle material.  This type of cracking in this type of
material has been identified at other facilities.  However, the cracks at Davis-Besse
appear to have initiated earlier than expected due to fabrication issues and plant
operating conditions.
*
The Davis-Besse staff, through their boric acid corrosion control program, did not clean
and inspect the reactor vessel head sufficiently to identify the leakage due to nozzle
cracking, nor the degradation of pressure boundary material.
*
The apparent rate of boric acid corrosion was consistent with certain industry data.
However, the corrosion rate used by the Babcock and Wilcox Owners Group, in their
past assessment of potential head degradation associated with nozzle cracking, was
significantly less than the apparent corrosion rate at Davis-Besse.
*
The Davis-Besse staff missed several opportunities to identify the boric acid corrosion of
the reactor vessel head at an earlier time.  These opportunities involved the failure to
identify the source of corrosion products that had accumulated on the containment air
cooler fins, deposited on the containment radiation element filters, and noted as
emanating from the inspection ports on the reactor vessel head service structure.
 
1
Report Details
1.0
BACKGROUND AND EVENT OVERVIEW
On March 6, 2002, Davis-Besse personnel notified the NRC of degradation to the
reactor vessel head material adjacent to a control rod drive nozzle.  The NRC issued a
Confirmatory Action Letter on March 13, 2002.  An Augmented Inspection Team (AIT)
was chartered in Attachment A to determine the facts and circumstances related to the
degradation of the reactor vessel head pressure boundary material, and to identify any
precursor indications of this condition.  The AIT developed a sequence of events,
interviewed plant personnel, collected and analyzed factual information and evidence
relevant to the reactor vessel head material loss, and conducted visual inspections of
the reactor vessel head.  The inspection was conducted in accordance with the AIT
Charter, NRC Inspection Procedure 93800, Augmented Inspection Team, and NRC
Management Directive 8.3, NRC Incident Investigation Program.  In accordance with
NRC procedures, the AIT charter did not include the verification of compliance with NRC
rules and regulations, the recommendation of enforcement actions, nor the
determination of risk significance for this issue.  A public exit was conducted on April 5,
2002, using the presentation material in Attachment B.
1.1
Description of Reactor Vessel Head and Penetration Nozzles
Davis-Besse Nuclear Power Station is a two-loop pressurized water reactor designed by
Babcock and Wilcox (B&W).  The Davis-Besse reactor vessel has a torispherical
shaped closure head constructed from low alloy steel (American Society of Mechanical
Engineers Boiler and Pressure Vessel Code (ASME Code), SA-533, Grade B, Class 1),
with approximately an 87-inch inside crown radius, 6.63 inches thick.  The inside surface
of the vessel head is clad with Type 308 and 308L stainless steel using a 6-wire
submerged arc welding process.  The cladding is provided for corrosion resistance and
is not credited as pressure boundary material. 
There are 69 vessel head penetration nozzles arranged in a rectangular pattern, with a
center-to-center distance of approximately 12 inches, and are numbered sequentially
starting at the center and progressing concentrically outward.  The nozzles are
fabricated from Alloy 600 tubes, with an outside diameter of approximately 4.00 inches
and a wall thickness of 0.65 inches.  The nozzles vary in length, depending on the
location on the vessel head, from approximately 30 inches in the center to approximately
50 inches on the periphery.  This includes a flange at the top for connecting to the
control rod drive mechanism (CRDM) housings.  Refer to Slide 5 in Attachment B for a
diagram of the CRDM configuration.  The nozzles extend through 4.00 inch bores in the
vessel head, and are welded to the head with a J-groove weld at the inner surface of the
head using Alloy 82 and 182 weld material.  Refer to Slide 7 in Attachment B for a
diagram of the CRDM nozzle.
The service structure is an enclosure attached to the reactor vessel head, approximately
18 feet high and 10 feet in diameter.  This structure stabilizes and houses the CRDMs
and contains a horizontal layer of metallic reflective insulation approximately 2 inches
above the top of the vessel head.  The CRDM nozzles welded to the vessel head pass
 
2
through the insulation layer and attach to the CRDM housings with bolted flanges.
These flanges are located about 9 inches above the horizontal insulation layer. 
1.2
Sequence of Events:  Discovery of Reactor Vessel Head Degradation
On February 16, 2002, the Davis-Besse facility began its 13th refueling outage (13 RFO),
which included inspections of the CRDM nozzles in accordance with NRC Bulletin
2001-01, "Circumferential Cracking of Reactor Pressure Vessel Head Penetration
Nozzles."  On February 27, 2002, the licensee notified the NRC that CRDM Nozzles 1, 2
and 3 exhibited axial through-wall indications.  The licensee decided to repair these
three nozzles plus two other nozzles which had crack indications that did not appear to
be through-wall.
On March 5, 2002, the licensee began repair work on CRDM Nozzle 3.  The repair
process included roll expansion of the CRDM nozzle material into the surrounding
reactor vessel head material, followed by machining along the axis of the CRDM nozzle
from the bottom to a point above the cracks in the nozzle material.  After machining up
past the J-groove weld, the machine unexpectedly rotated 15 degrees.  The machining
process was stopped and the machining tool was removed.  Subsequent investigation
identified that CRDM Nozzle 3 had tilted and was resting against an adjacent nozzle
flange, which indicated a loss of some vessel head material.
On March 6, 2002, the licensee began an investigation to identify the cause of the
movement by removing the CRDM nozzle.  At the same time, activities were underway
to remove boric acid residue from the top of the reactor vessel head using high pressure
hot water to dissolve the deposits.  After removing the boric acid deposits, the licensee
identified a large cavity in the head material on the downhill side of CRDM Nozzle 3.  In
addition, during this same time period, the licensee identified a smaller cavity in the
reactor vessel head after machining away the lower portion of Nozzle 2 during repair
activities.
2.0
CHARACTERIZATION OF NOZZLE CRACKING AND REACTOR VESSEL HEAD
WASTAGE AREAS
2.1
CRDM Nozzle Cracking
In response to NRC Bulletin 2001-01, Circumferential Cracking of Reactor Pressure
Vessel Head Penetration Nozzles, the licensee ultrasonically examined all 69 CRDM
nozzles during the current outage (13 RFO).  These examinations were conducted
inside the penetration tube from below the vessel head, and data was recorded from at
least 1 inch above the J-groove weld down to the lower end of the nozzle.  For these
examinations, the ultrasonic transducers used were mounted in a blade probe head and
setup for time-of-flight-diffraction.  The transducer orientation was such that it provided
maximum sensitivity for circumferentially oriented cracks near the outside diameter of
the tube.  Six nozzles were initially identified with crack-like indications using this
technique. 
 
3
For the six nozzles with crack indications a supplemental ultrasonic examination was
conducted using a rotating head probe from above the vessel head.  This probe head
contained several types and angles of transducers designed to maximize the response
to cracks oriented in both the circumferential and axial directions.  This rotating probe
confirmed cracks in five of the six nozzles identified by the blade probe.  The cracks in
these five nozzles initiated from the outside diameter of the nozzle near the J-groove
weld.  In three of the nozzles, through-wall axial cracks were identified that traversed the
J-groove weld area of the nozzle.  In addition, one circumferentially (circ.) oriented crack
was identified in Nozzle 2 just above the J-groove weld, that was about 50 percent
through-wall in depth.  The number and dimensions of nozzle cracks are identified
below:
Nozzle
Number
Cracks and
Orientation
Through-
Wall Cracks
Through-Wall Crack Length
(inches)
Crack Length Above
J-weld (inches)
1
9 Axial
2
1.77 and 3.49
0.0, 0.5
2
8 Axial
1 Circ.
5
None
3.86, 2.71, 2.59, 3.95, 3.04
Not Applicable
0.8, 0.5, 0.5, 1.0, 0.5
Not Applicable
3
4 Axial
2
4.08, 3.84
1.3, 0.8
5
1 Axial
None
Not Applicable
Not Applicable
47
1 Axial
None
Not Applicable
Not Applicable
Although cracking was not identified at Nozzle 46, ultrasonic examinations revealed
evidence of possible leakage and minor wastage in the annulus between the nozzle and
the vessel head.  Because a crack entirely within the J-groove weld could provide a
leakage path and would not be detected with ultrasonic techniques, the licensee
performed a dye penetrant examination of the J-groove weld.  Four rounded indications
were found, one 0.13 inches in diameter and three 0.06 inches in diameter.  At the
conclusion of this inspection, the licensee had not yet confirmed whether these
indications were indicative of J-groove weld cracking.
2.2
Reactor Vessel Head Wastage Areas
The cavity adjacent to Nozzle 3 extended downhill toward Nozzle 11 for approximately
5 to 7 inches and was 4 to 5 inches wide.  Within this area the 6.63 inch thick low
alloy steel head was corroded away leaving only the stainless steel cladding layer on
the inside of the reactor vessel head.  The remaining cladding layer, ranging in
thickness from 0.24 to 0.38 inches, had deflected upward into the cavity approximately
0.12 inches.  This cladding layer is designed as a corrosion resistant layer and no credit
is taken for the structural or pressure retaining capability of this layer.  Therefore, the
cavity at Nozzle 3 represented a loss of the design basis structural/pressure retaining
boundary for the vessel head. 
 
4
The cavity sides contained uneven ridges tapering downward, such that the cavity was
larger at the outer surface of the head.  Additionally, an undercut shelf existed at the
downhill end of the cavity near Nozzle 11.  An ultrasonic examination was conducted
from the inner surface of the head to determine the extent of the cavity near Nozzle 3.
This examination found that the cavity potentially had a debonding area between the
stainless steel cladding layer and the vessel head material which extended for several
inches around the cavity.  The licensee intended to conduct additional examinations to
further quantify the extent of this debonding.  Refer to Slides 8 and 9 of Attachment B
for a diagram and picture of this cavity. 
In addition to the cavity adjacent to Nozzle 3, a comparatively small cavity was identified
behind Nozzle 2.  This cavity was approximately 1.75 inches wide and 0.25 inches deep.
The licensee determined that the cavity extended from the top of the weld to the top of
the vessel behind Nozzle 2 (approximately 4.2 inches).  Refer to Slide 10 of
Attachment B for a diagram of this area.  The licensee removed Nozzle 2 to provide a
more detailed characterization of this cavity after the AIT inspection.
3.0
PROBABLE CAUSE OF NOZZLE CRACKING AND HEAD WASTAGE
3.1
Probable Cause for Nozzle Cracking
For the five penetration nozzles with indications characterized as cracks (Section 2.1),
four of these nozzles (Nos. 1, 2, 3, 5) were made from material heat No. M3935
manufactured by B&W Tubular Products.  This same heat of tube material was found
to have cracks in 14 of 68 penetrations used at Oconee Unit 3.  This cracking was
confirmed to be primary water stress corrosion cracking (PWSCC) based on analysis of
cracked nozzles removed from Oconee Units 2 and 3 (these units also have a vessel
head designed and constructed by B&W).  Therefore, based on the observed
susceptible heat of nozzle material under a similar environment, the AIT concluded that
the Davis-Besse nozzle cracking was likely caused by PWSCC.
3.1.1
Factors Affecting Primary Water Stress Corrosion Cracking of Nozzles
Cracking of Inconel Alloy 600 penetration nozzle materials near the J-groove weld has
been observed at several pressurized water reactors.  The area of the J-groove weld on
the nozzle is susceptible to PWSCC as discussed in NRC Generic Letter (GL) 97-01,
Degradation of Control Rod Drive Mechanism Nozzle and Other Vessel Closure Head
Penetrations, and in NRC Information Notice 2001-05, Circumferential Cracking of
Reactor Pressure Vessel Head Penetration Nozzles at Oconee Nuclear Station, Unit 3.
The susceptibility of a nozzle to cracking has been reviewed and documented in
NUREG/CR-6245, Assessment of Pressurized Water Reactor Control Rod Drive
Mechanism Nozzle Cracking.  The susceptibility of a nozzle to PWSCC may be
dependant on material, operating temperature, time, environment and residual stress.
Because the operating environment of domestic pressurized water reactors is similar,
the susceptibility of a particular nozzle to cracking may be dependant upon time,
temperature, material microstructure and residual tensile stress.  Thus, a particular heat
of Alloy 600 used to fabricate a penetration nozzle may be more likely to experience
cracking as each of these variables is increased (e.g., longer service time, higher
 
5
operating temperatures, or a higher residual tensile stress).  For the J-groove weld
connecting the nozzle to the vessel head, a high residual tensile hoop stress is
developed in the nozzle because of weld shrinkage.  The magnitude of this residual
tensile stress can range up to the yield strength of the material.
Crack initiation for PWSCC is strongly dependant on temperature (NUREG/CR-6245).
The 605oF operating temperature at Davis-Besse is higher than the other B&W plants
(typically 602oF).  This higher operating temperature may have shortened the required
operating time required to initiate cracking in the nozzles at Davis-Besse relative to
other B&W designed plants.
Once a crack is formed (at a given temperature and environment) in a nozzle, the speed
of crack propagation may be influenced by the tensile hoop stress induced from plant
operating pressure and residual tensile hoop stresses induced by welding.  As an axial
crack in the nozzle progresses in length above the J-groove weld, welding induced
residual tensile stress decreases rapidly, leaving only the operating pressure hoop
stresses to extend the crack length.  This results in slower crack growth as a crack
increases in length above the J-weld.  Therefore, the cracks identified in Section 2.1
which extend for the greatest distance above the J-groove weld are potentially the oldest
cracks.
3.1.2
CRDM Nozzle Materials and Contributing Factors
Of the 69 Alloy 600 nozzles at Davis-Besse, 60 were manufactured by B&W Tubular
Products and 9 were fabricated by Huntington Alloys.  The nozzles are attached to the
vessel head with an Alloy 82/182 butter and Alloy 82/182 J-groove weld.  The specific
method of fabricating the nozzle tubes was not recorded, but it would include rotary
piercing or extruding over a mandrel followed by a mill anneal.  The mill annealing heat
treatment temperature should be in the range of 1850oF to 1950oF to put carbon into
solution so that the carbides will precipitate at the grain boundaries during cooling.  This
heat treatment also redistributes chromium in the region of the grain boundaries.
However, based on review of production records, the nozzles for all B&W plants were
mill annealed in the temperature range of 1600oF to 1700oF.  This lower temperature
can increase susceptibility to primary water stress corrosion cracking.
As stated above, four of the Davis-Besse nozzles (Nos. 1, 2, 3, 5) exhibiting cracks were
fabricated from material heat No. M3935 manufactured by B&W Tubular Products.  This
nozzle material heat had the highest yield strength (48,500 pounds per square inch) of
the four material heats used to fabricate Davis-Besse head penetrations.  It appears that
this heat of Alloy 600 is more susceptible to primary water stress corrosion cracking
than other heats of Alloy 600 used for B&W penetration tubes.  However, the Owners
Groups for B&W, Westinghouse, and Combustion Engineering have not been able to
establish a definitive correlation between the yield strength and susceptibility to
primary water stress corrosion cracking.  Penetration tube 47 was also manufactured
by B&W Tubular Products (heat number C2649-1) and contained a small crack below
the J-Groove weld.  This heat of material had the second highest yield strength
(44,900 pounds per square inch).  An additional factor affecting the materials yield
stress was the straightening process used during manufacturing.  This process will work
 
6
harden the outside diameter of the nozzle resulting in the outside diameter yield stress
being substantially above inside diameter yield stress.
3.2
Probable Cause for Vessel Head Wastage Cavities
Corrosion experiments (discussed in Section 3.2.1.2) simulating a cracked nozzle have
confirmed that corrosion rates in excess of 2 inches per year are possible in low alloy
steel.  Nozzle 3 contained two through-wall axial cracks, which traversed the J-groove
weld.  The longest of these two cracks extended for approximately 1.3 inches above the
J-groove weld.  This crack would likely be the oldest crack in this nozzle as discussed in
Section 3.1.1.  The crack was on the downhill side of Nozzle 3 in direct alignment with
the long dimension of the cavity.  Therefore, the AIT concluded that the cavity observed
on Nozzle 3 was associated with boric acid corrosion from crack induced leakage at this
nozzle.  Further, the AIT concluded, based on corrosion products observed on the head
and in the containment air coolers and radiation element filters, that the corrosion
process had been in progress for at least 4 years.
For Nozzle 2, the crack with the longest dimension above the J-weld was also located in
the same area as the observed area of metal loss behind this nozzle.  Again, the AIT
considered that the metal loss was caused by boric acid corrosion from crack induced
leakage at this nozzle.
3.2.1
Boric Acid Corrosion Mechanism
Pressurized water reactors use boric acid in the reactor coolant as one means of
controlling the nuclear reaction rate.  The levels of boric acid in the reactor coolant can
range up to 2000 parts per million, which is generally not corrosive to materials used in
the reactor plant.  However, if boric acid is allowed to reach a concentrated solution it
can become very corrosive to carbon steel components.  The NRC issued GL 88-05,
Boric Acid Corrosion of Carbon Steel Reactor Pressure Boundary Components in PWR
[Pressurized Water Reactor] Plants, in March of 1988.  The Generic Letter was in
response to several industry incidents where concentrated boric acid solution, formed by
evaporation of water from leaking reactor coolant, corroded reactor coolant pressure
boundary components.  The Generic Letter requested that licensees implement a
program consisting of systematic measures to ensure that the reactor coolant pressure
boundary would have an extremely low probability of abnormal leakage, rapidly
propagating failure, or gross rupture. 
3.2.1.1 Boric Acid Corrosion Processes
Compounds of boron can develop from the precipitation of boric acid from solution.
Boric acid (H3BO3) and boric oxide (B2O3) can exist in a solid or molten state.  The solid
form of boric acid produced during evaporation depends on the rate of evaporation with
faster evaporation creating smaller particles.  When a boric acid solution comes in
contact with boric acid crystals, larger crystals tend to form.  It is also possible to form a
salt tree when previously precipitated solids form a porous structure that can wick more
solution to the vapor phase interface.
 
7
Boric acid solution that leaks onto the vessel head will cause the water to flash to steam,
leaving behind white, popcorn-like boric acid crystals.  This form of boric acid crystals is
relatively easy to remove after the reactor is cooled down to ambient temperature.  Dry,
white, powdery boric acid crystals on the reactor vessel head have been found to be
relatively benign while the reactor head is at operating temperatures.  Although some
darkening of the boric acid crystals may occur with age, brown or rust colored boric acid
is a strong indication that corrosion has occurred and a problem potentially exists.
Above 302oF, boric acid begins to dehydrate to form boric oxide:
2 H3BO3  B2O3 + 3H2O
The final condition of the mixture of boric acid and boric oxide is site specific, depending
on the relative quantities of each component and the amount of flow of boric acid, the
porosity created by steam escaping, and the presence of impurities such as iron oxide.
Boric oxide begins to soften at 617oF and becomes highly viscous at 842oF.
As boric acid that is not converted to the oxide is heated above 365oF, it may become
a viscous fluid (A. S. Myerson, Handbook of Industrial Crystallization, Butterworth-
Heinemann, Boston, 1993), conforming to the surrounding geometry under the influence
of gravity.  Molten boric acid can contain between 8 and 14 percent water and can be
highly corrosive under some conditions (U. Gurbuz Beker and N. Bulutcu, A New
Process to Produce Granular Boric Oxide by High Temperature Dehydration of Boric
Acid in a Fluidized Bed, Transactions of the Institute of Chemical Engineers, 74A, 133,
1996).  Discussions with the NRC staff and staff members at the Brookhaven National
Laboratory indicate that the boric acid/boric oxide mixture can vitrify if concentrated
sufficiently and held at a high enough temperature.
3.2.1.2 Industry Accepted Boric Acid Corrosion Rates
In GL 88-05 corrosion rates were identified for pressure boundary materials of up to
0.019 inches per year (in/yr) at 500oF.  For lower temperatures, corrosion rates up to
4.8 in/yr were identified.  However, these corrosion rates were established for
configurations which were not representative of the CRDM nozzle to head annulus gap
configuration.
A Babcock & Wilcox (B&W) owners group report, BAW-10190P, Safety Evaluation For
B&W Design Reactor Vessel Head Control Rod Drive Mechanism Nozzle Cracking,
was completed in May of 1993.  In this report, a Combustion Engineering pressurizer
heater sleeve mockup was used as the basis for establishing a 1.07 cubic inches per
year corrosion rate as the applicable rate for the vessel head due to cracks in CRDM
nozzles.  The test results used by B&W were documented in EPRI report TR-102748S,
Boric Acid Corrosion Guidebook.  This B&W analysis concluded that with this
corrosion rate, a plant would remain within ASME Code structural requirements for a
minimum of 6 years.  The AIT identified a test note in the EPRI report which stated that
the maximum volume loss of 1.07 cubic inches per year may not be conservative for all
cases since the volume loss is likely to increase as the corrosion depth and wetted
surface area increase.
 
8
In November of 2001, an EPRI test was documented in Revision 1 to the Boric Acid
Corrosion Guidebook.  This test was performed utilizing a configuration, temperature,
materials and leak rates which more closely matched the CRDM nozzle to vessel
configuration.  This test identified a corrosion rate of up to 2.37 in/yr.  This test also
indicated that the maximum corrosion occurred at the location where the boric acid
entered the annulus gap.  The contour of the degradation observed at Nozzle 2 and
Nozzle 3 appeared to support this test result.
3.2.2
Licensee Preliminary Identified Cause
The preliminary conclusions of the licensees root cause team were documented in a
memorandum to the Davis-Besse Site Vice President, dated March 22, 2002
(Attachment C).  In this memorandum, the root cause team concluded:  The factors
that caused corrosion of the reactor pressure vessel (RPV) head in the regions of
nozzles #2 and #3 are the CRDM nozzle leakage associated with through-wall cracking,
followed by boric acid corrosion of the RPV low-alloy steel.  The root cause team
concluded that the cracking initiated in Nozzle 3 in 1990 (+/- 3 years) and the crack had
propagated through-wall between 1994 and 1996.  The average rate of RPV head
corrosion was identified as 2 inches per year along the line from Nozzle 3 to Nozzle 11.
In this memorandum, the root cause team also stated that:  The estimated corrosion
rates are compatible with test results reported in Electric Power Research Institutes
(EPRI) Boric Acid Corrosion Guidebook.  They are also consistent with the video,
photographic and supporting plant data, that show that significant corrosion was
occurring by the 1998 to 1999 time-frame.  In addition, the root cause team identified a
number of causal factors such as boric acid accumulation on the top of the RPV head
and flange leakage.
The AIT concluded that the licensees root cause team had reviewed the applicable
historical data and established an appropriate time-line that supported the root cause.
Although the AIT agreed with the preliminary root cause conclusions, there were several
crucial questions left unanswered.  The licensees root cause efforts were continuing at
the conclusion of the NRCs inspection.  After the conclusion of the AIT, the licensee
provided their final root cause analysis report to the NRC, on April 18, 2002, and
provided responses to the NRCs questions associated with the preliminary root cause
report on April 30, 2002.  These documents are currently under review.
4.0
HISTORY OF VESSEL HEAD INSPECTIONS AND MATERIAL CONDITION
4.1
Background CRDM Flange Leakage
Historically, CRDM flange leakage had been observed at several B&W designed plants.
At Davis-Besse, CRDM flange leakage typically resulted in deposits of boric acid on the
service structure above the reflective insulation.  However, flange leakage in liquid form
also ran down the nozzles through the clearance gaps in the insulation and became
boric acid deposits on the vessel head.  The access for removing the boric acid deposits
and inspecting the vessel head for corrosion is through (18) 5-inch by 7-inch rectangular
openings or weep holes.  These openings are at the bottom of the service structure
 
9
where it is attached to the vessel head.  This location combined with the curvature of the
vessel head made it difficult to inspect and clean the top center portion of the vessel
head.  Visual inspections of the vessel head have typically been accomplished using
small video cameras inserted through the weep holes.  Refer to Slide 5 in Attachment B
for a diagram of the vessel head.
The CRDM flanges and flange bolts are made of stainless steel, corrosion resistant
materials.  Although the split nut-rings, located on the underside of the lower flange
face, are made of a low alloy steel and are susceptible to corrosion, they have been
coated with a corrosion resistant product.  The nut-rings have not been found with boric
acid corrosion at Davis-Besse.  Because of these corrosion resistant materials, leakage
from CRDM flanges typically does not result in corrosion, and any boric acid deposits
from flange leakage are normally white or light in color.  Conversely, as documented in
the Davis-Besse Boric Acid Corrosion Control Procedure, boric acid deposits with red or
rust color indicate that corrosion has occurred. 
The licensee systematically resolved CRDM flange leakage by replacing the flange
gaskets with a new design.  Starting in 6 RFO (1990), gaskets were replaced on flanges
which had developed leaks during the previous operating cycle, such that by 10 RFO
(1996), the last nine old-design gaskets were replaced even though these flanges were
not leaking.
4.2
History of Flange Leakage and Reactor Head Inspections
Inspections of the reactor head associated with identifying boric acid deposits were
recorded after the licensee established a Boric Acid Control Program in 1988 in
response to NRC GL 88-05.  The following inspection results were documented in the
licensees corrective action system through PCAQRs [potential conditions adverse to
quality reports] or CRs [condition reports] and/or recorded on video-tapes:
*
In April of 1990 (6 RFO) 22 leaking CRDM flanges were identified and repaired
(PCAQR 90-0120).
*
In September of 1991 (7 RFO) 15 out of 21 leaking CRDM flanges were
repaired.  Boric acid was observed on the reactor vessel head that ran along the
curvature of the head and stopped on the vessel closure bolts (PCAQR 91-
0353).  The source of these deposits was identified as flange leakage. Cleaning
was performed with a wire brush and vacuum.  No surface irregularities were
noted following cleaning; however, the extent of deposits if any that remained
after cleaning was not documented.
*
In March of 1993 (8 RFO) 14 leaking CRDM flanges were identified and 11 were
repaired (PCAQR 93-0132).  The boric acid from flange leakage was removed to
the extent possible by washdown of the head (PCAQR 96-551).  The AIT viewed
a videotape of the head inspection conducted during this outage and prior to the
head washdown.  Discrete patches of brown and white boric acid deposits were
observed which were more numerous near the center of the head.
 
10
*
In October of 1994 (9 RFO) eight CRDM flanges were leaking.  All eight were
repaired including three leaking flanges from the previous outage (PCAQR 94-
0912).  No record of a reactor vessel head inspection could be found.
*
In April of 1996 (10 RFO) the remaining nine CRDM flanges (non-leaking) not
previously repaired were modified with an enhanced gasket design.  The
head was inspected and video-taped using a remote camera mounted to a
hand-held pole inserted through the weep holes.  Several patches of boric
acid accumulation were identified including a brown stained deposit at
Nozzle 67 (PCAQR 96-551).  The licensee documented that boron deposits
could be indicative of flange leakage or nozzle leakage.  A vacuum was used to
remove boric acid deposits, but was not fully effective at removing the deposits
of boric acid near the center of the head.  The corrosion on the head from
remaining boric acid was evaluated and considered minimal based on B&W
Document 51-1229638, which identified minimal boric acid corrosion of carbon
steel head material at temperatures corresponding to the normal head operating
temperature.  The licensee concluded that 50 to 60 percent of the head had
been examined during this inspection.  The limited head examination appeared
to be due to access restrictions caused by the weep hole access limitations and
the curvature of the head.  The AIT observed the videotaped inspection and
noted that the boric acid deposits were generally white in color and appeared to
be the consistency of loose powder and discrete lumps.
*
In May of 1998 (11 RFO) one leaking CRDM flange was identified and not
repaired (PCAQR 98-0649).  The head was inspected and video-taped using a
remote camera mounted to a hand-held pole inserted through the weep holes.
This inspection identified areas near the center of the head covered with an
uneven layer of boric acid (PCAQR 98-767).  The licensee documented that the
boric acid deposits were removed as best as we can.  The boric acid color was
rust brown, which the licensee attributed to old deposits of boric acid.  The
previous root cause investigation and source documents from PCAQR 96-551
were referenced as the basis for leaving boric acid deposits on the head.  The
licensee concluded that due to the minimal operating time below 550F, there
was no impact on vessel head integrity.  Based on review of this video-taped
inspection, the AIT identified consolidated boric acid deposits near the center
region including Nozzle 2 and 3 locations.  On the head at an elevation below
Nozzles 3 and 11, the AIT noted that the boric acid appeared highly adherent
and rust brown in color.
*
In April of 2000 (12 RFO) five leaking CRDM flanges were identified and repaired
(CR 2000-0782).  The head was inspected and video-taped using a remote
camera mounted to a hand-held pole inserted through the weep holes.  Lava-
like brown/red deposits of boric acid over 1-inch thick were observed on much of
the vessel head (CR 2000-1037).  The corrective action for this condition was to
repeat cleaning of  the head until most of the boric acid deposits are removed.
Licensee logs recorded that crowbars were needed to remove the solid rock
hard deposits of boron on the head.  In addition, pressurized heated water was
used to remove the boric acid deposits.  The extent of remaining boric acid
deposits or evaluation of the effects on the head was not documented in the
 
11
corrective action system after this cleaning.  The system engineer also reported
a large amount of boric acid deposits were observed above the mirror insulation
due to flange leakage.  The AIT viewed the video-taped examination made with a
remote camera after the cleaning.  This videotape showed a thick layer of lava-
like brown/red boric acid that remained around the nozzles in the center of the
head.
*
In February of 2002 (13 RFO) no CRDM flange leakage was identified.
The head was inspected and video-taped using a remote camera mounted
to a hand-held pole inserted through the weep holes.  The licensee
documented that more boron than expected was found on the top of the head
(CR 02-00685).  Because the head was covered with boric acid and debris
deposits, indications of nozzle crack induced leakage could not be positively
identified at any nozzle location.  The AIT reviewed pictures and tapes of this
head inspection, which showed a thick lava-like brown/red deposit of boric acid
covering the center of the head.  Specifically, for 12 nozzles near the center of
the head, the boric acid layer was several inches thick and precluded access for
the remote camera inspection.  The licensee subsequently removed the boric
acid deposits from the head using hot pressurized water and identified the large
head cavity at Nozzle 3.
The AIT noted the following important aspects in the above history of inspections and
material condition of the RPV head:
(1)
No flange leakage was found during 10 RFO (1996), and very limited flange
leakage was noted during 11 RFO (1998).  However, boric acid accumulation on
the reactor vessel head increased from 9 RFO (1994) to 10 RFO (1996) and
from 10 RFO (1996) to 11RFO (1998).  Although the boric acid accumulation did
not come from flange leakage, the licensee apparently did not deduce that it then
must have come from pressure boundary leakage, such as nozzle cracking.
(2)
Although five flanges were documented as leaking during 12 RFO (2000),
according to CR-2000-0782, only four of the flanges showed positive evidence of
gasket leakage.  The fifth flange did not show the typical signs of flange leakage,
but boric acid deposits had built up under the flange to the extent that the flange
could not be fully inspected.  This flange was for Nozzle 3, and the licensee
concluded that the boric acid buildup was due to the flange leaking.  The
licensee apparently did not consider that the boric acid buildup could be due to
nozzle leakage from below.
(3)
Pictures of the reactor vessel, attached to CR-2000-0782, showed rust colored
boric acid deposits emanating from the inspection openings on the reactor
vessel head service structure.  Although the licensees boric acid corrosion
control procedure specifically stated that corrosion will most likely be exhibited by
rust stained boric acid, the source of these corrosion products was not
addressed in the condition report.
 
12
5.0
OPPORTUNITIES FOR EARLY DETECTION OF HEAD DEGRADATION
The AIT evaluated plant indications that could have provided an early opportunity to
detect the corrosion occurring in the vessel head.  The AIT identified the following
indicators which could have provided early detection of the head corrosion.
5.1
Boric Acid Corrosion Control Program
Leakage from the reactor coolant system (RCS) with the reactor at power will flash to
steam and leave behind boric acid crystals.  Averaged over the course of a fuel cycle,
there is approximately 0.03 pounds of boric acid per gallon of primary coolant.
Assuming a leak rate of 0.001 gallons per minute, approximately 15 pounds of boric acid
crystals would be produced in the vicinity of the vessel head by a postulated crack in a
CRDM nozzle over one year.  This leak rate would be significantly less than the
minimum detection capability of the plant leakage detection systems.  Therefore,
inspection of the reactor head for boric acid deposits is potentially the most sensitive
method available for detecting small leaks caused by cracked nozzles.  However, there
are limitations to this method.  First, depending on location, a leak may not be
accessible with the reactor at power.  Consequently, certain leaks can only be identified
when the reactor is shut down, which may only occur during refueling outages every two
years.  Second, this method depends on removing all existing boric acid accumulation,
so any new leak can be detected without being masked by previous accumulations.
This is critical because very small leaks may not be identifiable if the preexisting
accumulation is not removed.
As previously discussed in Section 4.2, the licensee had preformed visual inspections
of the reactor vessel head in 7 RFO (1991), and 8 RFO (1993) in accordance with
GL 88-05 guidance.  Davis-Besses implementing procedure for GL 88-05 was
NG-EN-00324, Boric Acid Corrosion Control.  Although recurring CRDM flange
leakage was documented during 9 RFO (1994), licensee personnel were unable to
identify any records documenting the visual inspections of the head during that outage.
In addition, boric acid deposits have historically been left on the head from flange
leakage as discussed in Section 4.  A leaking flange typically results in boric acid
deposits which travel down past the head insulation resulting in a deposit/buildup of
boric acid on the head.  In accordance with the boric acid control program, these
deposits should have been removed and the head inspected and any corrosion
evaluated.
During 10 RFO (April 1996), a licensee engineer initiated PCAQR 96-0551,
Boric Acid on Reactor Vessel Head, to document that the steps required by
Procedure NG-EN-00324, Boric Acid Corrosion Control, had not been followed during
the previous outage and that the procedure could not be fully implemented due to
limited access to the reactor vessel head.  The evaluation presented in this PCAQR
acknowledged the need to clean the vessel head, such that nozzle leakage could be
detected in the future.  Also, the initial assessment in this PCAQR stated that the failure
to clean the boric acid deposits made it difficult to determine if the deposits occurred
 
13
because of leaking flanges or because of a crack in the CRDM nozzle.  Licensee
managers approved the PCAQRs initial assessment subject to the following comment:
Nozzle cracking is of course a significant issue.  However, at present, the
probability of occurrence is relatively low.  We should remove boron from the
reactor pressure vessel head as best we can and so as to minimize dose.  This
will allow us to monitor any leakage, should a nozzle crack initiate.
The corrective action for this PCAQR became a Request for Modification 94-0025 (see
Section 5.5.1 for additional discussion on the delay of this modification). 
Because of access limitations (see Sections 4.1 and 4.2), the RPV head was not
completely cleaned and some portions were not thoroughly inspected, as specified by
the licensees Boric Acid Corrosion Control Program.  The bases for not cleaning or
inspecting the CRDM nozzles near the center of the RPV head was documented in
PCAQRs or provided by licensee staff during interviews with the AIT.  Specifically, the
following information was utilized by the licensee to justify leaving boric acid deposits on
the RPV head as identified during inspections in 10 RFO, 11 RFO and 12 RFO:
1) B&W Owners Group stress analyses had predicted that peripheral nozzles
were more likely to crack than nozzles near the center of the vessel head. 
2) Dried boric acid was not corrosive to the vessel and moderate amounts of
boric acid from CRDM flange leakage had historically been found and
cleaned up in the past, with no vessel corrosion.
3) Very limited boric acid corrosion occurs in the temperature range existing at
the vessel head.
4) EPRIs Boric Acid Corrosion Guidebook indicated that, under specific
circumstances, a layer of boric acid potentially protects a surface from
ongoing corrosion by keeping water away from the surface.
5) CRDM nozzle cracking was an age related phenomenon, and the
Davis-Besse staff believed they should not see any cracking because it was
several years younger than Oconee where significant problems had not yet
occurred.  This was codified by the B&W Owners Group in July 1997 through
a probabilistic susceptibility ranking that was developed in response to the
NRCs GL 97-01.
The identification of nozzle cracks at Oconee Units 1 & 3, prompted the NRC to issue
Bulletin 2001-01, which requested licensees to provide information, including a
description of their previous inspections of the reactor vessel head.  The Davis-Besse
responses of September 4 and October 17, 2001, described their previous inspection
and noted that, since 1996, four of the nozzles in the center of the vessel head were
obscured with boric acid deposits and could not be viewed.  In addition, the licensees
responses described their analytical efforts to verify that gaps would exist between the
CRDM nozzles and the reactor vessel head, permitting through-wall leakage from a
crack in a nozzle to be observed via boric acid deposits.
The licensees analyses concluded that, except for Nozzles 1, 2, 3, and 4 (center
nozzles), gaps would exist during normal operating conditions through which leakage
could occur and boric acid deposits would be evident.  In their supplemental response to
 
14
the NRC Bulletin, dated October 30, 2001, the licensee stated that based on the above
analytical results, the Davis-Besse staff would not expect to see boric acid residue
around Nozzles 1, 2, 3, or 4 if a crack were present.  This was based on the
manufactured interference fit between the nozzles and the vessel head.  The notable
aspect of this conclusion was that the analytically predicted interferences ranged from
0.000025 to 0.000004 inches.  Because the fabrication tolerances were more than an
order of magnitude greater than the analytical results, the AIT considered the licensees
conclusion, relative to not expecting boric acid residue if a crack were present in these
nozzles, to be unrealistic.
During interviews with the AIT, licensee personnel acknowledged that the reactor vessel
head was treated less rigorously than other components in the plant, within the context
of the GL 88-05 program.  Although the boric acid corrosion control program was
appropriately entered when boric acid was identified on the reactor vessel head, the
resolution of the issue was not treated the same.  Using the longstanding rationale
discussed above, the licensee used a philosophy that boric acid had been on the reactor
head for many years and no problems had ever been found.
5.2
Reactor Coolant System Leakage Detection
Because leakage from the through-wall cracks in Nozzle 3 would result in reactor
coolant leakage into the containment atmosphere, the leakage detection systems in
containment were reviewed to determine whether this system could have provided an
early indicator of head corrosion.  The observed leakage rate from a cracked nozzle
would be expected to be very small based on a leakage rate (0.003 gallons per minute
(gpm)) attributed to CRDM nozzle cracks observed at a foreign reactor plant (Bugey).
Regulatory Guide 1.45, Reactor Coolant Pressure Boundary Leakage Detection
Systems, details requirements for leakage monitoring equipment such as the
containment atmosphere particulate and gaseous radioactivity monitoring systems and
containment sump level/flow monitoring system.  The licensee has implemented a leak
detection program in accordance with Regulatory Guide 1.45 as described in the
Updated Safety Analysis Report, Section 5.2.4.
Reactor coolant system (RCS) leakage is grouped into two categories:  identified and
unidentified.  Identified leakage is that which is captured and metered through closed
systems, such as a collecting tank (e.g., pump seals and valve packing leaks); leakage
into containment atmosphere from sources that are both specifically located and known
not to interfere with the operation of leakage detection systems or not to be pressure
boundary leakage; leakage through the steam generators to the secondary system; and
reactor coolant pump seal returns.  Unidentified leakage is everything which is not
identified leakage. 
Unidentified RCS leakage was normally less than 0.1 gpm (monthly average), until
October of 1998, when a decision was made to remove the rupture disks downstream of
the pressurizer relief valves for design concerns (PCAQR 98-1980).  Specifically, a drain
line, designed to collect relief valve leakage in the quench tank, was bypassed in this
modification.  This allowed leakage past the relief valves to be vented directly into the
containment atmosphere, which collected in the normal sump and added to the
 
15
unidentified leakage, which increased to a maximum of 0.8 gpm.  During a mid-cycle
outage in May of 1999, the licensee resolved this design concern by installing new
rupture disks and reconnecting the drain line.  This resulted in a decrease in
unidentified leakage.  However, the unidentified leakage returned to levels between
0.15 and 0.25 gpm.  Subsequent investigations and containment entries were not
successful in identifying definitive sources of this leakage.  The licensee concluded,
based upon the history of CRDM flange leakage and that unidentified leakage values
observed at Davis-Besse were near industry averages, the leakage was most likely from
the CRDM flanges.
Because of historical variations in unidentified leakage compared to the relatively small
amount of leakage associated with CRDM cracks, the AIT concluded that, by itself,
unidentified leakage trends were not a reasonable method of detecting nozzle cracking.
However, when considered together with other indications of corrosion products as
discussed in Section 4.2, above, and in Sections 5.3 and 5.4 below, the AIT concluded
that this was a missed opportunity to detect the corrosion occurring on the reactor
vessel head.
5.3
Containment Air Coolers
Reactor coolant leakage through the cracks in Nozzle 3 would travel as steam and liquid
in the annulus behind the nozzle and leave boric acid deposits on the top of the head.
In addition, this steam leakage would cause boric acid and corrosion products from the
head cavity to be divided into fine particles which would be dispersed into the air space
above the head.  These fine particles would then be captured by the service structure
ventilation system intake and be distributed throughout the containment.  A key area
which could collect these airborne particles of boric acid and corrosion products is at the
containment air coolers (CAC).
The vessel head service structure ventilation pulls a suction from the CRDM flange
area through the fans located on the 603 feet elevation, exhausting through ductwork
to the top of the East D-ring.  This provided a potential pathway for any corrosion fines
and boric acid particulate dispersion originating from the vessel head.  In November of
2001, radiological surveys showed a contamination plume effect originating from the
service structure ventilation exhaust over the East D-ring.  However, an isotopic analysis
was not performed of the plume to fully characterize the source of the contamination.
Additionally, two containment recirculation fans provide a mixing of the containment
atmosphere, further dispersing the fines and particulates.
The CAC system consists of three separate tube/fin coolers (which are cooled by the
service water (SW) system) located inside containment, and connected to a common
supply plenum.  Downstream of this plenum is a ductwork distribution system, designed
to distribute air over and around all heat producing equipment, such as the reactor
vessel, D-rings (housing the steam generators, pressurizer and reactor coolant pumps)
and incore instrument tank.  The external surfaces of the cooler tube banks are readily
visible from the outside of the coolers, and have a remote indication of plenum pressure
(used to determine cooling fin fouling) in the control room.
 
16
If a leak occurs from the RCS during normal operations, an aerosol mist is produced
from the water flashing and evaporating as it exits the leak, increasing containment
ambient humidity.  Since the inlet water temperature of SW to the CACs is normally
between 40°F and 75°F, substantially cooler than containment air temperatures, the
CACs condense this ambient humidity to water, which is ultimately collected in the
normal containment sump.  In the process of removing the humidity, the CACs also
collect particulate boric acid (which would be released with the RCS leakage as fine
particles) on the cooling fins, in the discharge plenum and the associated ductwork.
This fouling will decrease the plenum pressure, as read remotely in the control room,
during periods of high boric acid accumulation.
In 1992, the licensee had experienced a CAC fouling from a leak in the reactor head
vent line flange to the primary side of the steam generator.  As a result, the licensee
cleaned the boric acid, evident by the uniformly white coating on all three coolers.  After
repairs to the flange, no further boric-acid precipitated cleanings were required for
several years.
In October of 1998, the removal of the rupture disks downstream of the pressurizer relief
valves substantially contributed to the RCS unidentified leakage.  In November 1998,
PCAQR 98-1980 identified that the CAC fouling had increased correspondingly to
increased leakage from the pressurizer reliefs.  The CACs were cleaned 17 times from
November 1998 to May 1999.  During a mid-cycle outage in May 1999, the design
concern was resolved, the rupture disks reinstalled, and the drain line reconnected.
However, two additional CAC cleanings were conducted, one in June 1999 and one
in July 1999.  The post-job critique observed the boric acid to be rust color on and
in the boron being cleaned away from CAC No. 1.  Subsequent interviews indicated
this was presumed to be the result of restoring from the mid-cycle outage, and the
residual humidity in containment from outage-related repairs.  After being cleaned in
July 1999, the CACs did not need any further cleaning for approximately 10 months.
Although the licensee installed high efficiency particulate air filters (inside containment)
during August and September 1999, this did not appear to factor into the need for CAC
cleaning.
After 12 RFO (May of 2000), CAC deposits were again forming, as evidenced by the
decrease in plenum pressure.  Eight CAC cleanings were conducted between
June 2000 and May 2001, with no further cleanings required through the end of cycle.
However, for 13 RFO (February 2002), the licensee reported (15) 5-gallon buckets of
boric acid were removed from the ductwork and plenum.  Significant boric acid was
found elsewhere within containment, including on SW piping, stairwells and other areas
of low ventilation.
After the 1999 mid-cycle outage, the licensee had attributed the excessive boric acid
accumulation and CAC cleanings to leakage from CRDM flanges.  In 12 RFO
(May 2000), several leaking flanges were repaired, the results of which could not be
verified throughout the cycle.  However, 13 RFO (February 2002) inspections indicated
the repairs had been successful, and no flange leakage was detected.  Furthermore,
earlier experience with leaking flanges (pre-1992, and 1992-1998) did not result in the
need to clean the CACs.  Therefore, CRDM flange leakage would not have reasonably
been the major contributor to the increased boric acid loading on the CACs during this
 
17
time frame.  The licensee had also attributed the discoloration of the boric acid to
migration of the surface corrosion on the CACs into the boric acid and the aging of the
boric acid itself.
The AIT considered the sudden change to rust colored boric acid deposits in June of
1999, to indicate corrosion product accumulation from the formation of the head cavity
near Nozzle 3.  The failure of the licensee to identify the source of these deposits
represented a missed opportunity to identify the corrosion cavity in the head at that time.
5.4
Radiation Elements
As discussed in Section 5.2, steam leakage through the cracks in Nozzle 3 would result
in fine particles of boric acid and corrosion products.  These particles would then be
captured by the service structure ventilation system intake and distributed throughout
the containment.  An area where these fine particles of boric acid and corrosion
products would be collected and observed is in the radiation element (RE) system filters.
There are two identical radiation element air sampling systems, drawing from two
sample locations within containment.  Air samples are drawn from within containment,
passed through a particulate filter, an iodine sample cartridge and a noble gas detector
before being exhausted back into containment.  Both systems normally draw a sample
from near the top of the D-ring structures, but can also draw from near the polar crane,
and near the personnel airlock on the 603 feet elevation.
Boric acid accumulation on the RE filters can clog the filters and decrease flow to below
acceptable levels, necessitating a filter change.  Licensee records correlate past RCS
leakage increases with RE filter changes, such as in 1992 when the reactor head vent
flange leakage caused this to occur.  In March of 1999, RE filter clogging from boric acid
deposits was attributed to the pressurizer relief valve rupture disk maintenance which
occurred in 1998.  Filter changes normally occurred based on a monthly schedule rather
than low flow rates.  Beginning in May of 1999, the schedule of filter change out went
from a monthly interval to an irregular 1 to 3 week interval, occasionally dropping to a 1
to 2 day interval by November 1999.  In response to the increased frequency of filter
changeouts, the licensee installed two large high efficiency particulate air filter units
inside containment to capture a large portion of the corrosion fines.  Additionally, the RE
sample points were changed to the alternate locations.  This action appeared to improve
the service life of the filters, but did not eliminate the filter loading conditions completely.
In May of 1999, the RE filters began accumulating a yellowish-brown material.  This
material was sent to an external laboratory for analysis.  The results of this analysis
were received in November 1999, and positively identified the presence of ferric oxide.
Specifically, this analysis stated, The fineness of the iron oxide (assumed to be ferric
oxide) particulate would indicate it probably was formed from a very small steam leak.
The particulate was likely originally ferrous hydroxide in small condensed droplets of
steam and was oxidized to ferric oxide in the air before it settled on the filters; and the
iron oxide does not appear to be coming from the general corrosion of a bare metal
surface in containment or from steam impingement on a metal surface. 
 
18
Accumulation of boric acid on the RE filters was readily recognized as a symptom of
RCS leakage.  During 12 RFO, CRDM flange D10 was attributed as the source of the
RCS leakage, since the flange required machining to correct the leakage.  However, the
presence of ferric oxide fines was not explained, nor were multiple containment entries
successful in determining a source.  Additionally, past CRDM flange leakage had not
significantly contributed to the CAC fouling, nor the RCS leakage indications.
Therefore, the AIT believed that the corrosion deposits first identified in the RE filters
beginning in May of 1999, indicated that corrosion was occurring due to the formation of
the head cavity near Nozzle 3.  The failure of the licensee to identify the source of these
corrosion products represented a missed opportunity to identify the corrosion cavity in
the head at that time.
5.5
Causal Factors Influencing Head Degradation Detection
Several decisions made by Davis-Besse personnel at various times directly influenced
or potentially affected their ability to detect the head degradation associated with the
CRDM nozzle leakage. These are discussed below.
5.5.1
Decision to Delay Modification to Service Structure
In March of 1990, modification 90-0012 was initiated to install multiple access ports in
the service structure to permit inspection and cleaning of the vessel head.  This
modification was canceled in 1992, because the current inspection techniques were
considered adequate.
In March of 1994, a licensee engineer initiated PCAQR 94-0295 to question why there
was no commitment requiring a visual inspection of the reactor vessel head every
refueling outage, as referenced in the NRC 1993 Safety Evaluation for the Alloy 600
CRDM nozzle cracking issue.  The PCAQRs response from the Nuclear Assurance
Director indicated that the commitment for the visual inspection did not appear to have
been a licensee commitment to the NRC.  Regulatory Affairs and Design Engineering
personnel indicated that, although an enhanced visual was not a commitment to the
NRC, they recommended the visual inspection be done.  However, the plant engineering
staffs comment in the PCAQR stated that there was a low risk of a crack in CRDM
nozzles since none had been identified in the United States, and that the available
inspection methods were not highly reliable.  On that basis, the plant engineer felt it was
not necessary to perform the inspections.
In May of 1994, the licensee engineer who wrote the above PCAQR initiated a Request
for Modification (RFM 94-0025) to install openings in the CRDM service structure to
allow thorough inspection and cleaning of the reactor vessel head.  The modification
request noted that, out of all of the B&W plants, only Davis-Besse and Arkansas
Nuclear One, Unit 1, had not installed the access openings in the service structure.  The
modification request cited the following reasons for the modification: 
1) there was no access to the reactor vessel head or CRDM nozzles without
the modification, and there was an ongoing industry concern for Alloy 600
nozzle cracking;
 
19
2) inspection of the reactor vessel head for boric acid corrosion was difficult
and not always adequate, because the video inspections did not encompass
a 100 percent inspection of the head;
3) cleaning boric acid residue from the vessel head did not encompass
100 percent, because the size and geometry of the weep holes only
permitted cleaning of the lower one-third of the head with scrapers and wire
brushes. 
The modification was approved by the plant in July of 1994, but remained unfunded by
the Project Review Committee/Project Review Group until November of 1998, when it
was scheduled for implementation in 13 RFO (2002).  The modification was
subsequently deferred until 14 RFO by the Project Review Group, as part of an effort
to meet the 2001/2002 expenditure targets by reducing the number of projects
implemented.  In discussing the reasons for not implementing this modification, the
rationale identified in Section 5.1 were also applied.  The AIT considered the delay in
implementing the modification as  contributing to the failure to detect head degradation.
5.5.2
Decision to Delay Repair of CRDM Flange on Nozzle 31 in 11RFO
During 8 RFO (1993), CRDM flange leakage was noted on several CRDM flanges
including the flange for Nozzle 31.  The corrective actions included polishing the flange
surface and replacing the gasket with a new design.  The PCAQR issued to document
this condition (93-0132) contained a recommendation that the flange surface be
inspected during each subsequent maintenance outage and be machined if further
leakage occurs.  During 11 RFO (1998), the CRDM flange for Nozzle 31 was found to
be leaking, and as indicated in PCAQR 98-0649, the amount of leakage was not
considered significant compared to flange leakage from previous outages.
Consequently, no corrective actions were taken, even though the vendor (Framatome)
reiterated their recommendation from 1993 to machine the flange.  The PCAQR did
contain a recommendation to reexamine the flange for Nozzle 31 during 12 RFO and to
replace the gasket if the flange was leaking.
During 12 RFO (2000), significant flange leakage was noted and five leaking flanges
were identified during the video inspections of the CRDM flanges, including Nozzle 31s.
The majority of the boric acid accumulation was attributed to Nozzle 31s flange due to
steam cutting of the flange face.  Condition Report 2000-1037 was written to describe
the boric acid accumulation on the RPV head and on top of the insulation.  The boric
acid accumulation was attributed to leaking CRD flanges.  The AIT considered the delay
in repairing Nozzle 31s flange as a contributing cause of this event, because the
extensive amount of flange leakage contributed to the boric acid deposits on the head
which masked evidence of the nozzle leakage occurring at this time.
6.0
CONCLUSIONS
The AIT presented the inspection results to Mr. Saunders and other members of the
licensee management at the conclusion of the inspection on April 5, 2002.  The licensee
acknowledged the conclusions presented as discussed in Attachment B and
summarized below.
 
20
The AIT concluded that the probable cause of the cavity at Nozzle 3 was boric acid
corrosion of the head associated with reactor coolant leakage from a through-wall crack
in this nozzle.  Further, the AIT concluded based on corrosion products observed on the
head, and in the CAC and RE filters that the corrosion process had been in progress for
at least 4 years.
The AIT concluded that the probable cause of the cracking observed in the five
penetration nozzles was PWSCC.  This was based on similar cracking identified at two
other B&W plants that performed destructive analysis of cracked nozzles fabricated
from the same heat of material to confirm PWSCC.
The AIT evaluated the indications which existed that could have provided an early
opportunity to detect evidence of the formation of the corrosion cavity in the head at
Nozzle 3.  The AIT identified several opportunities which were available to the licensee
to potentially identify this corrosion cavity at an earlier point in time.  Specifically, these
missed opportunities were associated with the failure to identify the source of the
corrosion products deposited in the CAC and RE filters in early 1999 and the failure to
remove boric acid or evaluate the source of corrosion products which accumulated on
the vessel head.
 
21
KEY POINTS OF CONTACT
DAVIS-BESSE
H. Bergendahl, Vice President - Nuclear
D. Eshelman, Director, Support Services
R. Fast, Plant Manager
D. Geisen, Manager, Design Engineering
D. Lockwood, Manager, Regulatory Affairs
J. Messina, Director, Work Management
D. Miller, Supervisor, Compliance
S. Moffit, Director, Technical Services
R. Saunders, President, FirstEnergy Nuclear Operating Company
NUCLEAR REGULATORY COMMISSION
J. Davis, Sr. Material Engineer, NRR
R. Gardner, Chief, Engineering Branch
J. Grobe, Director, Division of Reactor Safety
C. Lipa, Chief, Reactor Projects Branch 4
B. Sheron, Associate Director for Project Licensing and Technical Analysis, NRR
LIST OF ACRONYMS USED
AIT
Augmented Inspection Team
ASME
American Society of Mechanical Engineers
B&W
Babcock and Wilcox
CAC
Containment Air Cooler
CR
Condition Report
CRDM
Control Rod Drive Mechanism
EPRI
Electric Power Research Institute
GL
Generic Letter
gpm
Gallon Per Minute
in/yr
Inches Per Year
NRC
Nuclear Regulatory Commission
PCAQR
Potential Conditions Adverse to Quality Report
PDR
Public Document Room
PWSCC
Primary Water Stress Corrosion Cracking
RCS
Reactor Coolant System
RE
Radiation Element
RFO
Refueling Outage
RPV
Reactor Pressure Vessel
SW
Service Water
 
22
LIST OF DOCUMENTS REVIEWED
Calculation
SIA Calc W-ENTP-11Q-306
Finite Element Gap Analysis of CRDM Penetrations
(Davis-Besse), October 8, 2001.
Condition Reports (CR)
1992-0139
Boron Found on Containment Air Sample Filter
1993-0187
Boric Acid Accumulation on SW Piping
1998-0020
Multiple Problems Identified with RC-2
1998-0330
Industry Event (Prairie Island) Crack in the Motor Tube of the Control Rod Drives
1998-1963
Design Over-Stress of the Pressurizer Nozzles for Safety Valve
1999-0372
Received Computer PT-RE4597AA/AB High
1999-0510
Low Flow Alarm Observed on RE4597BA While Out of Service for Maintenance
1999-0745
Small Clumps of Boric Acid Present on Wall Opposite of DH108
1999-0861
RE4597AA Sample Lines Were Found to be Full of Water
1999-0928
Increased Frequency of Particulate and Charcoal Filters for RE 4597BA Being
Changed
1999-0998
Awareness of Approaching the Tech Spec Limit for Maximum Ctmt Air Temp
1999-1300
Analysis of CTMT Radiation Monitor Filters
1999-1614
Due Date of LER Commitment Missed: Boric Acid Control Program Procedure
Change
2000-0781
Leakage from CRD Structure Blocked Visual Exam of Reactor Vessel Head
Studs
2000-0782
Inspection of Reactor Flange Indicated Boric Acid Leakage From Weep Holes
2000-0903
Two of 40 CRDM Hold Down Bolts Had Indications Found During VT-1
Inspection
2000-0994
RV Head CRDM Nozzle at Location F-10 has Large Pit in Outer Gasket Groove
2000-0995
RV Head CRDM Nozzle Flange at Location D-10 has Extensive Pitting Across
the Outer Gasket Groove. Inner Gasket Also Has Pitting
2000-1037
Inspection of Reactor Head Indicated Accumulation of Boron in Area of the CRD
Nozzle Penetration
2000-1210
During Installation of Control Rod Drive Assembly at Location D-10, on the
Reactor Head, it was Discovered that Top of Motor Tube for this Drive was out of
Line with Surrounding Motor Tubes
2000-1547
CAC Plennum Pressure Drop Following 12 RFO
2000-4138
Frequency for Cleaning Boron From CAC Fins Increased to Interval of
Approximately 8 weeks
2001-0039
CAC Plenum Pressure Experienced Step Drop
2001-0487
Certain Areas Inside CTMT in Year 2000 Seeing Higher Temperatures
2001-0890
Unidentified RCS Leak Rate Varies Daily by as Much as 100 percent of the
Value
2001-1110
Chemistry is Changing Filters on RE4597BA More Frequently
2001-1822
Frequency of Filter Changes for RE4597BA is Increasing
2001-1857
RCS Unidentified Leakage at .125 to .145 gpm
2001-2012
NRC Issuance of IEB 01-01 Circumferential Cracking of RX Pressure Vessel
Head Penetration Nozzles
 
23
2001-2769
RE2387 Identified Spiked Above ALERT and High Setpoints
2001-2795
RE4597BA Alarmed on Saturation
2001-2862
Calculated Unidentified Leakage for Reactor Coolant System has Indicated
Increasing Trend
2001-2936
Monthly Functional Test for RE4597BA/BB Count Not Performed
2001-3025
Increase in RCS Unidentified Leakage
2001-3411
Received Equipment Fail Alarm for Detector Saturation on RE4597BA
2002-0685
Loose Boron 1-2" deep 75% Around Circumference of Flange
2002-0846
More Boron Than Expected Found on Top of Head
2002-0891
UT Performed on #3 CRDM Nozzle Revealed Indication of Through-Wall Axial
Flaws
2002-0932
Completion of UT on All 69 CRDM Nozzles Revealed Additional CRDM Cracks
Beyond #3 Nozzle
2002-1053
While Machining Reactor Vessel Head Nozzle #3 the Nozzle Machining Tool
Moved Approximately 15 Degrees
2002-1128
Evaluation of Bottom up Ultrasonic Test Data in Area of RX Pressure Vessel
Head Nozzle #3 Shows Significant Degradation of RX Vessel Head Pressure
Boundary
2002-1159
During Video Tape Review, Indication Found on Newly Machined Face on Mid-
Span of CRDM Nozzle.  Appears to be Through-wall in Immediate Vicinity of
Base Metal Indications.
Drawings
M-503-127-3
Closure Head Assembly, Revision 3
M-503-212-1
Closure Head Subassembly Drawing, Revision 1
M-503-213-2
Closure Head Subassembly Drawing, Revision 2
03-1221681-03
Framatome Drawing of RV Nozzle/Nur Ring Modification
Modifications
MOD 90-0012
Modification Reactor Closure Head Access Ports
MOD 94-0025
Install Service Structure Inspection Openings
TM 1998-0036
Temporary Modification: Preliminary Evaluation of Pressurizer Nozzles for
Relief Valves Demonstrates that an Overstress Condition May Exist in the
Nozzle Flange
NRC Generic Communications for Control of Boric Acid Corrosion
IN 80-27
Degradation of Reactor Coolant Pump Studs, dated June 11, 1980
IEB 82-02
Degradation of Threaded Fasteners in the Reactor Coolant Pressure
Boundary of PWR Plants, dated June 2, 1982
IN 82-06
Failure of Steam Generator Primary Side Manway Closure Studs, dated
March 12, 1982
IN 86-108
Degradation of Reactor Coolant System Pressure Boundary Resulting from
Boric Acid Corrosion, dated December 29, 1986
IN 86-108
Supplement 1, dated April 20,1987
IN 86-108
Supplement 2, dated November 19, 1987
IN 86-108
Supplement 3, dated January 5, 1995
 
24
GL 88-05
Boric Acid Corrosion of Carbon Steel Reactor Pressure Boundary
Components in PWR Plants, dated March,17, 1988
IN 90-10
Primary Water Stress Corrosion Cracking (PWSCC) of Inconel 600, dated
February 23,1990
IN 94-63
Boric Acid Corrosion of Charging Pump Casing Caused by Cladding Cracks,
dated August 30, 1994
IN 96-11
Ingress of Demineralizer Resins Increases Potential for Stress Corrosion
Cracking of Control Rod Drive Mechanism Penetrations, dated
February 14,1996
GL 97-01
Degradation of CRDM/CEDM Nozzle and other Vessel Closure Head
Penetrations, dated April 1, 1997
IN 2001-05
Through-wall Circumferential Cracks of Reactor Pressure Vessel Head
Control Rod Drive Mechanism Penetration Nozzles at Oconee Nuclear
Station, Unit 3, dated April 30, 2001
Bulletin 2001-01
Circumferential Cracking of Reactor Pressure Vessel Head Penetration
Nozzles, dated August 3, 2001
Other Documents
RAS02-00132
Probable Cause Summary Report for CR2002-0891, dated March 22, 2002
NPE-96-00260
Control Rod Drive Nozzle Cracking, dated May 8, 1996
Books
RCS System Performance Books Volumes 1 though 11
BAW-10190P
Safety Evaluation For B&W Design Reactor Vessel Head Control Rod Drive
Mechanism Nozzle Cracking, dated May of 1993
BAW-10190P,
Addendum 1
B&W Owners Group Proprietary, External Circumferential Crack Growth
Analysis for B&W-Design Reactor Vessel Head Control Rod Drive
Mechanism Nozzle Cracking, dated December 1993
BAW-2301
B&W Owners Group Proprietary, B&WOG Integrated Response to Generic
Letter 97-01, dated July 1997
Exam Report
Reactor Vessel Head ID Clad Thickness Measurements in Region of
Wastage Between Nozzles 3 and 11, dated March 18, 2002
Exam Report
DB-5
CRDM Nozzle 46 J-Groove Weld, dated March 24, 2002
Examination Report
Davis Besse 13 RFO CRDM Nozzle Examination Report, dated
March 11, 2002
Framatome
51-5015818-00
Davis-Besse CRDM Nozzle Heat Information, 2002
EPRI Report
TR-102748s
Boric Acid Corrosion Guidebook, Revision 0, dated April 1995
EPRI Report
1000975
Boric Acid Corrosion Guidebook, Revision 1, dated November 2001
Report 2779
Oconee Unit 3 CRDM Nozzle Crack and Material Characterization - Oconee
Unit 1 Thermocouple Tube material Characterization - Metallurgical Analysis
Report
Dominion
Engineering Report
Volume and Weight of Material Lost at Nozzle 3
51-125825-00
CRDM Nozzle Heat Treatment
 
25
Material Test
Report
DBNPS Reactor Vessel Head Certified Material Test Report
Intra-Company
Memorandum
Control Rod Drive Nozzle Cracking, dated May 8, 1996
Root Cause Plan Dated March 18, 2002.
Intra-Company
Memorandum
Probable Cause Summary Report for CR2002-0891, dated March 22, 2002
Meeting Minutes
DBPRC Meeting Minutes for MOD 94-0025.
Standing Order
87-015
RCS Leakage Management and Attached Policy Reactor Coolant System
Leakage Management
0620-00143210
Lukens Steel Company, Test Certificate, Chemical Analysis and Physical
Properties
Photographic Records
Picture
Vessel Head on Stand
Picture
Head and Service Structure Looking NE
Picture 
Scaffolding Around Service Structure
Picture
View From Newly Cut Service Structure Manway Opening Looking into Drives
Picture
Looking Through Manway Cut in Service Structure
Picture
View of CRD Flanges Above Insulation Showing Some Removed and Some
Installed
Picture
Control Rod Drive Flanges Above Insulation
Pictures
Shielded Work Platform on Top of Service Structure
Picture
Pictures of Nozzle 2 and 3
Pictures
Area Surrounding Nozzle 3 Penetration
Picture
Nozzle 16 Quad C
Pictures
Nozzle 2
Pictures
Nozzle 3 Remnant
Pictures
From Bare Head Video Exam Conducted in 13 RFO.
Video Tape
Davis-Besse Reactor Head Inspection Under Insulation Alloy 600, 12 RFO
Video Tape
Davis-Besse 12 RFO Final Head Inspection
Video Tape
Davis-Besse Reactor Head Cleaning 11 RFO
Video Tape
Davis-Besse Weep Hole Cleaning Nozzle 67, 10 RFO
Video Tape
Davis-Besse Weep Hole Video Inspection 10 RFO
Video Tape
13 RFO Reactor Head Nozzle Remote Visual Inspection
Video Tape
Root Cause Video of Nozzle #3 and Adjacent Nozzles, March 13, 2002 to
March 14, 2002
Video Tape
PT of Nozzle #46 J-groove Weld, March 24, 2002
Potential Conditions Adverse to Quality Reports (PCAQR)
1988-0494
Condition Not Satisfactorily Resolved per PCAQ
1990-0221
CRDM Flanges #F02 and F-4 Erosion and Irregularities.
1991-0353
Boron on Reactor Vessel Head
1992-0072
CAC Cleaning
1993-0098
Boric Acid Corrosion on OTSGA Head Vent Flange
1993-0132
Reactor Coolant Leakage from CRD Flange
 
26
1994-0912
Documents Results of CRDM leakage Video Inspection
1994-0974
CRDM Flange Indication
1994-0975
CRDM  Flange Indication
1994-1338
10 CFR Part 21 RX Adaptor Tubes
1996-0551
Boric Acid on RX Vessel Head
1996-0650
VT-2 Inspection Revealed Evidence of Leakage and Boric Acid Residue
1996-1018
IN 96-032 RV Augmented ISI
1998-0020
Inadequate Testing
1998-0649
Reactor Vessel Head Boron Deposits
1998-0767
Reactor Vessel Head Inspection Results
1998-0824
CAC Boric Acid Accumulation
1998-1164
Water in RE4597 Sample Lines
1998-1885
Found Two Carbon Steel Nuts on RC2
1998-1895
CTMT Normal Sump Leakage in Excess of 1 gpm
1998-1965
Water and Boron Accumulation on Filter Cartridges
1998-1980
Potential CAC Fouling
1998-2071
Accumulation of Boric Acid on CTMT Service Water Piping
Procedures
NG-EN-00324
Boric Acid Corrosion Control, Revisions 1, 2, and 3
PP-1102.10
Surveillance Test Procedure:  Plant Shutdown and Cooldown, Revision 16
DB-OP-06903
Operations Procedure:  Plant Shutdown and Cooldown
DB-PF-00204
ASME Section XI Pressure Testing, Revision 3
DB-OP-01200
Reactor Coolant System Leakage Management, Revision 3
 
ATTACHMENT A TO NRC AUGMENTED INSPECTION REPORT NO. 50-346/02-03(DRS)
USNRC Memorandum from J. E. Dyer to R. N. Gardner, dated March 12, 2002:  Augmented
Inspection Team Charter - Davis-Besse Reactor Vessel Head Material Loss
Documented in ADAMS (Accession Number ML020730194)
 
ATTACHMENT B TO NRC AUGMENTED INSPECTION REPORT NO. 50-346/02-03(DRS)
NRC Briefing Slides for the Public AIT Exit Meeting Conducted on April 5, 2002
Documented in ADAMS (Accession Number ML021070811).
 
ATTACHMENT C TO NRC AUGMENTED INSPECTION REPORT NO. 50-346/02-03(DRS)
FirstEnergy Intra-Company Memorandum from S. A. Loehlein to H. W. Bergendahl, dated
March 22, 2002
Documented in ADAMS (Accession Number ML020860035)
 
March 12, 2002
MEMORANDUM TO: Ronald N. Gardner, Chief
Electrical Engineering Branch
Division of Reactor Safety
FROM:
J. E. Dyer /RA/
Regional Administrator
SUBJECT:
AUGMENTED INSPECTION TEAM CHARTER -
DAVIS BESSE REACTOR VESSEL HEAD MATERIAL LOSS
In response to preliminary information provided by the licensee on March 10, 2002, regarding
the significant loss of pressure boundary material from the reactor vessel head, an augmented
inspection team (AIT) is being sent to the Davis-Besse Plant.  You are hereby designated as
the AIT leader.
A.
Basis
On March 6, 2002, during repair activities to control rod drive mechanism (CRDM)
nozzles, the licensee identified an area of wastage in the reactor pressure vessel head
surrounding the No. 3 CRDM nozzle.  The licensee initially identified five CRDM nozzles
that required repairs due to cracking in the J-groove welds found during the nozzle
examinations required by Bulletin 2001-01.  Wastage area in the head was discovered
when the licensee removed the No. 3 CRDM nozzle, after the penetration tube
unexpectedly moved during repair activities. 
Because this was a significant unplanned degraded condition having potential generic
safety implications, an AIT was initiated in accordance with NRC Management Directive
8.3, "NRC Incident Investigation Program."  The purpose of the AIT is to better
understand the facts and circumstances related to the degradation of the reactor vessel
head pressure boundary material.  It is also to identify any precursor indications of this
condition so that appropriate followup actions can be taken.  All followup actions
associated with the extent of condition, repairs/replacements, or corrective actions
related to plant restart will be covered through other inspection activities.
CONTACT:
John A. Grobe, Director, DRS
(630) 829-9700
 
R. Gardner
-2-
B.
Scope
Specifically, the augmented inspection team is expected to collect, analyze, and
document factual information and evidence sufficiently to address the following:
1.
The plant history of reactor coolant system operational leakage indications,
including trends in unidentified leakage, containment air cooler fouling,
containment radiation monitor readings, etc.
2.
The plant history of reactor vessel head material condition issues, including
control rod drive flange leakage or other sources of corrosive substances.
3.
The plant history of reactor vessel head inspection, including visual inspections,
ultrasonic testing, prior video-records of head examinations, reactor vessel head
cleaning activities, and licensee action in response to generic correspondence
for leakage and degradation of the reactor coolant system.
4.
Characterization of all reactor vessel head wastage areas, including the best
available geometric details of cavity volumes, surface conditions, surface
contaminants, etc.
5.
The probable cause(s) for the vessel head wastage.
C.
Guidance
This memorandum designates you as the AIT leader.  Your duties will be as described
in Inspection Procedure 93800, "Augmented Inspection Team."  The team composition
has been discussed with you directly.  During performance of the augmented inspection,
designated team members are separated from their normal duties and report directly to
you.  The team is to emphasize fact-finding in its review of the circumstances
surrounding the event, and it is not the responsibility of the team to examine the
regulatory process, to determine whether NRC requirements were violated, to address
licensee actions related to plant restart, or to address the applicability of generic safety
concerns to other facilities.  Safety concerns identified that are not directly related to the
event should be reported to the Region III office for appropriate action.
The team will report to the site, conduct an entrance meeting, and begin inspection on
Tuesday, March 12, 2002.  Tentatively, the inspection should be completed by close of
business March 22, 2002, with a report documenting the results of the inspection,
including findings and conclusions, issued within 30 days of the public exit meeting.
While the team is on site, you will provide daily status briefings to Region III
management.
This Charter may be modified should the team develop significant new information that
warrants review.
 
DOCUMENT NAME:  G:DRS\\ML021260141.wpd 
To receive a copy of this document, indicate in the box:  "C" = Copy without attachment/enclosure  "E" = Copy with attachment/enclosure  "N" = No copy
OFFICE
RIII
RIII
RIII
RIII
NAME
JGavula:sd
CLipa   
JGrobe       
JDyer
DATE
3/12/02
3/12/02
3/12/02
3/12/02
OFFICIAL RECORD COPY
}}

Latest revision as of 18:37, 16 January 2025

IR 05000346/2002-003, Davis-Besse Nuclear Power Station. Inspection Conducted on 03/12-04/05/2002, Augumented Inspection Team
ML021260141
Person / Time
Site: Davis Besse Cleveland Electric icon.png
Issue date: 05/03/2002
From: Dyer J
NRC/RGN-III/DNMS/FCB
To: Bergendahl H
FirstEnergy Nuclear Operating Co
References
IR-02-003
Download: ML021260141 (41)
v • d • e


See also: IR 05000346/2002003

Text

May 3, 2002

Mr. Howard Bergendahl

Vice President - Nuclear

FirstEnergy Nuclear Operating Company

Davis-Besse Nuclear Power Station

5501 North State Route 2

Oak Harbor, OH 43449-9760

SUBJECT:

DAVIS-BESSE NUCLEAR POWER STATION

NRC AUGMENTED INSPECTION TEAM - DEGRADATION OF THE

REACTOR PRESSURE VESSEL HEAD - REPORT NO. 50-346/02-03(DRS)

Dear Mr. Bergendahl:

Your staff provided information to the NRC between March 6 and 10, 2002, concerning the

identification of a large cavity in the reactor vessel head adjacent to a control rod drive nozzle.

On March 13, 2002, the NRC issued a Confirmatory Action Letter outlining specific actions your

staff are expected to take in response to this event. One of those actions is obtaining NRC

approval prior to restart of the Davis-Besse plant.

On March 12, 2002, the NRC dispatched an Augmented Inspection Team (AIT) to the Davis-

Besse site in accordance with NRC Management Directive 8.3, NRC Incident Investigation

Program. The AIT was chartered to determine the facts and circumstances related to the

significant degradation of the reactor vessel head pressure boundary material. The AIT

developed a sequence of events, interviewed plant personnel, collected and analyzed factual

information relevant to the degraded condition and conducted visual inspections of the reactor

vessel head. The enclosed report provides the AIT findings which were summarized for you

and your staff during a public exit meeting on April 5, 2002.

The cavity in the reactor vessel head was discovered during maintenance activities for

problems found during inspections conducted pursuant to NRC Bulletin 2001-01,

Circumferential Cracking of Reactor Pressure Vessel Head Penetration Nozzles. The

degraded area covers approximately 30 square inches where the thick low-alloy structural steel

was corroded away, leaving only the thin stainless steel cladding layer as a pressure boundary

for the reactor coolant system. This represents a loss of the reactor vessels pressure retaining

design function, since the cladding was not considered as pressure boundary material in the

structural design of the reactor pressure vessel. While the cladding did provide a pressure

retaining capability during reactor operations, the identified degradation represents an

unacceptable reduction in the margin of safety of one of the three principal fission product

barriers at the Davis-Besse Nuclear Power Station.

The AIT concluded that the cavity was caused by boric acid corrosion from leaks through the

control rod drive nozzles in the reactor vessel. These leaks were caused by primary water

stress corrosion cracking of the nozzle material leading to a through-wall crack and corrosion of

low alloy steel that went undetected for an extended period of time. The boric acid corrosion

H. Bergendahl

-2-

control program at the site included both cleaning and inspection requirements, but was not

effectively implemented to detect the leakage and prevent the significant corrosion of the

reactor vessel head over a period of years. Similarly on several occasions, maintenance and

corrective action activities failed to detect and address the indications in the containment that

the significant corrosion of the reactor vessel head was occurring. The NRC views these as

missed opportunities to identify and correct this significant degradation to the reactor pressure

vessel head.

The AIT did not address the verification of compliance with NRC rules and regulations, provide

recommendations for enforcement actions, or assess the risk significance of this issue. A

followup special inspection effort will be scheduled in the near future to pursue these aspects of

the regulatory process.

In accordance with 10 CFR 2.790 of the NRCs "Rules of Practice," a copy of this letter and

its enclosures will be available electronically for public inspection in the NRC Public

Document Room or from the Publicly Available Records (PARS) component of NRCs

document system (ADAMS). ADAMS is accessible from the NRC Web site at

http://www.nrc.gov/NRC/ADAMS/index.html (the Public Electronic Reading Room).

Sincerely,

/RA by J. L. Caldwell for/

J. E. Dyer

Regional Administrator

Enclosure:

NRC Augmented Inspection Report

No. 50-346/02-03(DRS)

cc w/encl:

B. Saunders, President - FENOC

Plant Manager

Manager - Regulatory Affairs

M. OReilly, FirstEnergy

Ohio State Liaison Officer

R. Owen, Ohio Department of Health

Public Utilities Commission of Ohio

H. Bergendahl

-2-

control program at the site included both cleaning and inspection requirements, but was not

effectively implemented to detect the leakage and prevent the significant corrosion of the

reactor vessel head over a period of years. Similarly on several occasions, maintenance and

corrective action activities failed to detect and address the indications in the containment that

the significant corrosion of the reactor vessel head was occurring. The NRC views these as

missed opportunities to identify and correct this significant degradation to the reactor pressure

vessel head.

The AIT did not address the verification of compliance with NRC rules and regulations, provide

recommendations for enforcement actions or assess the risk significance of this issue. A

followup special inspection effort will be scheduled in the near future to pursue these aspects of

the regulatory process.

In accordance with 10 CFR 2.790 of the NRCs "Rules of Practice," a copy of this letter and

its enclosures will be available electronically for public inspection in the NRC Public

Document Room or from the Publicly Available Records (PARS) component of NRCs

document system (ADAMS). ADAMS is accessible from the NRC Web site at

http://www.nrc.gov/NRC/ADAMS/index.html (the Public Electronic Reading Room).

Sincerely,

J. E. Dyer

Regional Administrator

Enclosure:

NRC Augmented Inspection Report

No. 50-346/02-03(DRS)

cc w/encl:

B. Saunders, President - FENOC

Plant Manager

Manager - Regulatory Affairs

M. OReilly, FirstEnergy

Ohio State Liaison Officer

R. Owen, Ohio Department of Health

Public Utilities Commission of Ohio

DOCUMENT NAME: G:DRS\\ML021260141.wpd

To receive a copy of this document, indicate in the box: "C" = Copy without attachment/enclosure "E" = Copy with attachment/enclosure "N" = No copy

OFFICE

RIII

RIII

RIII

RIII

RIII

NAME

JGrobe for

RGardner:sd

CLipa

JStrasma

JGrobe

JCaldwell

for JDyer

DATE

05/03/02

05/03/02

05/03/02

05/03/02

05/

OFFICIAL RECORD COPY

H. Bergendahl

-3-

ADAMS Distribution:

Chairman Meserve

Commissioner Dicus

Commissioner Diaz

Commissioner McGaffigan

Commissioner Merrifield

W. Travers, EDO

W. Kane, OEDO

H. Nieh, OEDO

OCIO

ACRS

S. Collins, NRR

B. Sheron, NRR

J. Zwolinski, NRR

J. Strosnider, NRR

G. Holahan, NRR

B. Bateman, NRR

S. Bajwa, NRR

S. Long, NRR

OPA

OCA

H. Miller, RI

L. Reyes, RII

E. Merschoff, RIV

AJM

DFT

SPS1

RidsNrrDipmIipb

HBC

C. Ariano (hard copy)

DRPIII

DRSIII

PLB1

JRK1

U.S. NUCLEAR REGULATORY COMMISSION

REGION III

Docket No:

50-346

License No:

NPF-3

Report No:

50-346/02-03

Licensee:

FirstEnergy Nuclear Operating Company

Facility:

Davis-Besse Nuclear Power Station

Location:

5501 North State Route 2

Oak Harbor, OH 43449

Dates:

March 12 - April 5, 2002

Team Leader:

R. Gardner, Engineering Branch Chief, DRS

Inspectors:

J. Davis, Sr. Materials Engineer, RES

M. Holmberg, Sr. Reactor Inspector, DRS

J. Gavula, Sr. Reactor Inspector, DRS

D. Simpkins, Resident Inspector, Davis-Besse, DRP

Approved by:

John A. Grobe, Director

Division of Reactor Safety

ii

TABLE OF CONTENTS

PAGE

SUMMARY OF FINDINGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

1.0

BACKGROUND AND EVENT OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1

Description of Reactor Vessel Head and Penetration Nozzles . . . . . . . . . . . . . . 1

1.2

Sequence of Events: Discovery of Reactor Vessel Head Degradation . . . . . . . 2

2.0

CHARACTERIZATION OF NOZZLE CRACKING AND REACTOR VESSEL HEAD

WASTAGE AREAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2.1

CRDM Nozzle Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2.2

Reactor Vessel Head Wastage Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

3.0

PROBABLE CAUSE OF NOZZLE CRACKING AND HEAD WASTAGE . . . . . . . . . . . . 4

3.1

Probable Cause for Nozzle Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3.1.1

Factors Affecting Primary Water Stress Corrosion Cracking of Nozzles . 4

3.1.2

CRDM Nozzle Materials and Contributing Factors . . . . . . . . . . . . . . . . . 5

3.2

Probable Cause for Vessel Head Wastage Cavities . . . . . . . . . . . . . . . . . . . . . 6

3.2.1

Boric Acid Corrosion Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

3.2.1.1

Boric Acid Corrosion Processes . . . . . . . . . . . . . . . . . . . . . . . . . 6

3.2.1.2

Industry Accepted Boric Acid Corrosion Rates . . . . . . . . . . . . . . 7

3.2.2

Licensee Preliminary Identified Cause . . . . . . . . . . . . . . . . . . . . . . . . . . 8

4.0

HISTORY OF VESSEL HEAD INSPECTIONS AND MATERIAL CONDITION . . . . . . . 8

4.1

Background CRDM Flange Leakage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

4.2

History of Flange Leakage and Reactor Head Inspections

. . . . . . . . . . . . . . . . 9

5.0

OPPORTUNITIES FOR EARLY DETECTION OF HEAD DEGRADATION . . . . . . . . . 12

5.1

Boric Acid Corrosion Control Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

5.2

Reactor Coolant System Leakage Detection . . . . . . . . . . . . . . . . . . . . . . . . . . 14

5.3

Containment Air Coolers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

5.4

Radiation Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

5.5

Causal Factors Influencing Head Degradation Detection . . . . . . . . . . . . . . . . . 18

5.5.1

Decision to Delay Modification to Service Structure . . . . . . . . . . . . . . . 18

5.5.2

Decision to Delay Repair of CRDM Flange on Nozzle 31 in 11RFO . . . 19

6.0

CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

ATTACHMENTS

Attachment A -

USNRC Memorandum from J. E. Dyer to R. N. Gardner, dated March 12,

2002: Augmented Inspection Team Charter - Davis-Besse Reactor Vessel

Head Material Loss

Attachment B -

NRC Briefing Slides for the Public AIT Exit Meeting Conducted April 5, 2002

Attachment C -

FirstEnergy Intra-Company Memorandum from S. A. Loehlein to H. W.

Bergendahl, dated March 22, 2002

iii

SUMMARY OF FINDINGS

IR 05000346-02-03, on 03/12-04/05/2002, FirstEnergy Nuclear Operating Company,

Davis-Besse Nuclear Power Station. Augmented Inspection Team.

This report covers a 3-week inspection by an NRC Augmented Inspection Team for the

substantial loss of material from the reactor pressure vessel head.

On March 5 and 6, 2002, workers at Davis-Besse were repairing control rod drive

penetration Nozzle 3, following the identification of cracks detected through inspections

performed pursuant to NRC Bulletin 2002-01, Circumferential Cracking of Reactor

Pressure Vessel Head Penetration Nozzles. The workers discovered a large cavity, a

significant loss of metal adjacent to the control rod drive nozzle in the reactor vessel

head, that apparently resulted from boric acid corrosion of the reactor vessel head due

to leakage from the cracks in Nozzle 3.

The cracks in the control rod drive nozzles were apparently due to primary water stress

corrosion cracking of the Alloy 600 nozzle material. This type of cracking in this type of

material has been identified at other facilities. However, the cracks at Davis-Besse

appear to have initiated earlier than expected due to fabrication issues and plant

operating conditions.

The Davis-Besse staff, through their boric acid corrosion control program, did not clean

and inspect the reactor vessel head sufficiently to identify the leakage due to nozzle

cracking, nor the degradation of pressure boundary material.

The apparent rate of boric acid corrosion was consistent with certain industry data.

However, the corrosion rate used by the Babcock and Wilcox Owners Group, in their

past assessment of potential head degradation associated with nozzle cracking, was

significantly less than the apparent corrosion rate at Davis-Besse.

The Davis-Besse staff missed several opportunities to identify the boric acid corrosion of

the reactor vessel head at an earlier time. These opportunities involved the failure to

identify the source of corrosion products that had accumulated on the containment air

cooler fins, deposited on the containment radiation element filters, and noted as

emanating from the inspection ports on the reactor vessel head service structure.

1

Report Details

1.0

BACKGROUND AND EVENT OVERVIEW

On March 6, 2002, Davis-Besse personnel notified the NRC of degradation to the

reactor vessel head material adjacent to a control rod drive nozzle. The NRC issued a

Confirmatory Action Letter on March 13, 2002. An Augmented Inspection Team (AIT)

was chartered in Attachment A to determine the facts and circumstances related to the

degradation of the reactor vessel head pressure boundary material, and to identify any

precursor indications of this condition. The AIT developed a sequence of events,

interviewed plant personnel, collected and analyzed factual information and evidence

relevant to the reactor vessel head material loss, and conducted visual inspections of

the reactor vessel head. The inspection was conducted in accordance with the AIT

Charter, NRC Inspection Procedure 93800, Augmented Inspection Team, and NRC

Management Directive 8.3, NRC Incident Investigation Program. In accordance with

NRC procedures, the AIT charter did not include the verification of compliance with NRC

rules and regulations, the recommendation of enforcement actions, nor the

determination of risk significance for this issue. A public exit was conducted on April 5,

2002, using the presentation material in Attachment B.

1.1

Description of Reactor Vessel Head and Penetration Nozzles

Davis-Besse Nuclear Power Station is a two-loop pressurized water reactor designed by

Babcock and Wilcox (B&W). The Davis-Besse reactor vessel has a torispherical

shaped closure head constructed from low alloy steel (American Society of Mechanical

Engineers Boiler and Pressure Vessel Code (ASME Code), SA-533, Grade B, Class 1),

with approximately an 87-inch inside crown radius, 6.63 inches thick. The inside surface

of the vessel head is clad with Type 308 and 308L stainless steel using a 6-wire

submerged arc welding process. The cladding is provided for corrosion resistance and

is not credited as pressure boundary material.

There are 69 vessel head penetration nozzles arranged in a rectangular pattern, with a

center-to-center distance of approximately 12 inches, and are numbered sequentially

starting at the center and progressing concentrically outward. The nozzles are

fabricated from Alloy 600 tubes, with an outside diameter of approximately 4.00 inches

and a wall thickness of 0.65 inches. The nozzles vary in length, depending on the

location on the vessel head, from approximately 30 inches in the center to approximately

50 inches on the periphery. This includes a flange at the top for connecting to the

control rod drive mechanism (CRDM) housings. Refer to Slide 5 in Attachment B for a

diagram of the CRDM configuration. The nozzles extend through 4.00 inch bores in the

vessel head, and are welded to the head with a J-groove weld at the inner surface of the

head using Alloy 82 and 182 weld material. Refer to Slide 7 in Attachment B for a

diagram of the CRDM nozzle.

The service structure is an enclosure attached to the reactor vessel head, approximately

18 feet high and 10 feet in diameter. This structure stabilizes and houses the CRDMs

and contains a horizontal layer of metallic reflective insulation approximately 2 inches

above the top of the vessel head. The CRDM nozzles welded to the vessel head pass

2

through the insulation layer and attach to the CRDM housings with bolted flanges.

These flanges are located about 9 inches above the horizontal insulation layer.

1.2

Sequence of Events: Discovery of Reactor Vessel Head Degradation

On February 16, 2002, the Davis-Besse facility began its 13th refueling outage (13 RFO),

which included inspections of the CRDM nozzles in accordance with NRC Bulletin 2001-01, "Circumferential Cracking of Reactor Pressure Vessel Head Penetration

Nozzles." On February 27, 2002, the licensee notified the NRC that CRDM Nozzles 1, 2

and 3 exhibited axial through-wall indications. The licensee decided to repair these

three nozzles plus two other nozzles which had crack indications that did not appear to

be through-wall.

On March 5, 2002, the licensee began repair work on CRDM Nozzle 3. The repair

process included roll expansion of the CRDM nozzle material into the surrounding

reactor vessel head material, followed by machining along the axis of the CRDM nozzle

from the bottom to a point above the cracks in the nozzle material. After machining up

past the J-groove weld, the machine unexpectedly rotated 15 degrees. The machining

process was stopped and the machining tool was removed. Subsequent investigation

identified that CRDM Nozzle 3 had tilted and was resting against an adjacent nozzle

flange, which indicated a loss of some vessel head material.

On March 6, 2002, the licensee began an investigation to identify the cause of the

movement by removing the CRDM nozzle. At the same time, activities were underway

to remove boric acid residue from the top of the reactor vessel head using high pressure

hot water to dissolve the deposits. After removing the boric acid deposits, the licensee

identified a large cavity in the head material on the downhill side of CRDM Nozzle 3. In

addition, during this same time period, the licensee identified a smaller cavity in the

reactor vessel head after machining away the lower portion of Nozzle 2 during repair

activities.

2.0

CHARACTERIZATION OF NOZZLE CRACKING AND REACTOR VESSEL HEAD

WASTAGE AREAS

2.1

CRDM Nozzle Cracking

In response to NRC Bulletin 2001-01, Circumferential Cracking of Reactor Pressure

Vessel Head Penetration Nozzles, the licensee ultrasonically examined all 69 CRDM

nozzles during the current outage (13 RFO). These examinations were conducted

inside the penetration tube from below the vessel head, and data was recorded from at

least 1 inch above the J-groove weld down to the lower end of the nozzle. For these

examinations, the ultrasonic transducers used were mounted in a blade probe head and

setup for time-of-flight-diffraction. The transducer orientation was such that it provided

maximum sensitivity for circumferentially oriented cracks near the outside diameter of

the tube. Six nozzles were initially identified with crack-like indications using this

technique.

3

For the six nozzles with crack indications a supplemental ultrasonic examination was

conducted using a rotating head probe from above the vessel head. This probe head

contained several types and angles of transducers designed to maximize the response

to cracks oriented in both the circumferential and axial directions. This rotating probe

confirmed cracks in five of the six nozzles identified by the blade probe. The cracks in

these five nozzles initiated from the outside diameter of the nozzle near the J-groove

weld. In three of the nozzles, through-wall axial cracks were identified that traversed the

J-groove weld area of the nozzle. In addition, one circumferentially (circ.) oriented crack

was identified in Nozzle 2 just above the J-groove weld, that was about 50 percent

through-wall in depth. The number and dimensions of nozzle cracks are identified

below:

Nozzle

Number

Cracks and

Orientation

Through-

Wall Cracks

Through-Wall Crack Length

(inches)

Crack Length Above

J-weld (inches)

1

9 Axial

2

1.77 and 3.49

0.0, 0.5

2

8 Axial

1 Circ.

5

None

3.86, 2.71, 2.59, 3.95, 3.04

Not Applicable

0.8, 0.5, 0.5, 1.0, 0.5

Not Applicable

3

4 Axial

2

4.08, 3.84

1.3, 0.8

5

1 Axial

None

Not Applicable

Not Applicable

47

1 Axial

None

Not Applicable

Not Applicable

Although cracking was not identified at Nozzle 46, ultrasonic examinations revealed

evidence of possible leakage and minor wastage in the annulus between the nozzle and

the vessel head. Because a crack entirely within the J-groove weld could provide a

leakage path and would not be detected with ultrasonic techniques, the licensee

performed a dye penetrant examination of the J-groove weld. Four rounded indications

were found, one 0.13 inches in diameter and three 0.06 inches in diameter. At the

conclusion of this inspection, the licensee had not yet confirmed whether these

indications were indicative of J-groove weld cracking.

2.2

Reactor Vessel Head Wastage Areas

The cavity adjacent to Nozzle 3 extended downhill toward Nozzle 11 for approximately

5 to 7 inches and was 4 to 5 inches wide. Within this area the 6.63 inch thick low

alloy steel head was corroded away leaving only the stainless steel cladding layer on

the inside of the reactor vessel head. The remaining cladding layer, ranging in

thickness from 0.24 to 0.38 inches, had deflected upward into the cavity approximately

0.12 inches. This cladding layer is designed as a corrosion resistant layer and no credit

is taken for the structural or pressure retaining capability of this layer. Therefore, the

cavity at Nozzle 3 represented a loss of the design basis structural/pressure retaining

boundary for the vessel head.

4

The cavity sides contained uneven ridges tapering downward, such that the cavity was

larger at the outer surface of the head. Additionally, an undercut shelf existed at the

downhill end of the cavity near Nozzle 11. An ultrasonic examination was conducted

from the inner surface of the head to determine the extent of the cavity near Nozzle 3.

This examination found that the cavity potentially had a debonding area between the

stainless steel cladding layer and the vessel head material which extended for several

inches around the cavity. The licensee intended to conduct additional examinations to

further quantify the extent of this debonding. Refer to Slides 8 and 9 of Attachment B

for a diagram and picture of this cavity.

In addition to the cavity adjacent to Nozzle 3, a comparatively small cavity was identified

behind Nozzle 2. This cavity was approximately 1.75 inches wide and 0.25 inches deep.

The licensee determined that the cavity extended from the top of the weld to the top of

the vessel behind Nozzle 2 (approximately 4.2 inches). Refer to Slide 10 of

Attachment B for a diagram of this area. The licensee removed Nozzle 2 to provide a

more detailed characterization of this cavity after the AIT inspection.

3.0

PROBABLE CAUSE OF NOZZLE CRACKING AND HEAD WASTAGE

3.1

Probable Cause for Nozzle Cracking

For the five penetration nozzles with indications characterized as cracks (Section 2.1),

four of these nozzles (Nos. 1, 2, 3, 5) were made from material heat No. M3935

manufactured by B&W Tubular Products. This same heat of tube material was found

to have cracks in 14 of 68 penetrations used at Oconee Unit 3. This cracking was

confirmed to be primary water stress corrosion cracking (PWSCC) based on analysis of

cracked nozzles removed from Oconee Units 2 and 3 (these units also have a vessel

head designed and constructed by B&W). Therefore, based on the observed

susceptible heat of nozzle material under a similar environment, the AIT concluded that

the Davis-Besse nozzle cracking was likely caused by PWSCC.

3.1.1

Factors Affecting Primary Water Stress Corrosion Cracking of Nozzles

Cracking of Inconel Alloy 600 penetration nozzle materials near the J-groove weld has

been observed at several pressurized water reactors. The area of the J-groove weld on

the nozzle is susceptible to PWSCC as discussed in NRC Generic Letter (GL) 97-01,

Degradation of Control Rod Drive Mechanism Nozzle and Other Vessel Closure Head

Penetrations, and in NRC Information Notice 2001-05, Circumferential Cracking of

Reactor Pressure Vessel Head Penetration Nozzles at Oconee Nuclear Station, Unit 3.

The susceptibility of a nozzle to cracking has been reviewed and documented in

NUREG/CR-6245, Assessment of Pressurized Water Reactor Control Rod Drive

Mechanism Nozzle Cracking. The susceptibility of a nozzle to PWSCC may be

dependant on material, operating temperature, time, environment and residual stress.

Because the operating environment of domestic pressurized water reactors is similar,

the susceptibility of a particular nozzle to cracking may be dependant upon time,

temperature, material microstructure and residual tensile stress. Thus, a particular heat

of Alloy 600 used to fabricate a penetration nozzle may be more likely to experience

cracking as each of these variables is increased (e.g., longer service time, higher

5

operating temperatures, or a higher residual tensile stress). For the J-groove weld

connecting the nozzle to the vessel head, a high residual tensile hoop stress is

developed in the nozzle because of weld shrinkage. The magnitude of this residual

tensile stress can range up to the yield strength of the material.

Crack initiation for PWSCC is strongly dependant on temperature (NUREG/CR-6245).

The 605oF operating temperature at Davis-Besse is higher than the other B&W plants

(typically 602oF). This higher operating temperature may have shortened the required

operating time required to initiate cracking in the nozzles at Davis-Besse relative to

other B&W designed plants.

Once a crack is formed (at a given temperature and environment) in a nozzle, the speed

of crack propagation may be influenced by the tensile hoop stress induced from plant

operating pressure and residual tensile hoop stresses induced by welding. As an axial

crack in the nozzle progresses in length above the J-groove weld, welding induced

residual tensile stress decreases rapidly, leaving only the operating pressure hoop

stresses to extend the crack length. This results in slower crack growth as a crack

increases in length above the J-weld. Therefore, the cracks identified in Section 2.1

which extend for the greatest distance above the J-groove weld are potentially the oldest

cracks.

3.1.2

CRDM Nozzle Materials and Contributing Factors

Of the 69 Alloy 600 nozzles at Davis-Besse, 60 were manufactured by B&W Tubular

Products and 9 were fabricated by Huntington Alloys. The nozzles are attached to the

vessel head with an Alloy 82/182 butter and Alloy 82/182 J-groove weld. The specific

method of fabricating the nozzle tubes was not recorded, but it would include rotary

piercing or extruding over a mandrel followed by a mill anneal. The mill annealing heat

treatment temperature should be in the range of 1850oF to 1950oF to put carbon into

solution so that the carbides will precipitate at the grain boundaries during cooling. This

heat treatment also redistributes chromium in the region of the grain boundaries.

However, based on review of production records, the nozzles for all B&W plants were

mill annealed in the temperature range of 1600oF to 1700oF. This lower temperature

can increase susceptibility to primary water stress corrosion cracking.

As stated above, four of the Davis-Besse nozzles (Nos. 1, 2, 3, 5) exhibiting cracks were

fabricated from material heat No. M3935 manufactured by B&W Tubular Products. This

nozzle material heat had the highest yield strength (48,500 pounds per square inch) of

the four material heats used to fabricate Davis-Besse head penetrations. It appears that

this heat of Alloy 600 is more susceptible to primary water stress corrosion cracking

than other heats of Alloy 600 used for B&W penetration tubes. However, the Owners

Groups for B&W, Westinghouse, and Combustion Engineering have not been able to

establish a definitive correlation between the yield strength and susceptibility to

primary water stress corrosion cracking. Penetration tube 47 was also manufactured

by B&W Tubular Products (heat number C2649-1) and contained a small crack below

the J-Groove weld. This heat of material had the second highest yield strength

(44,900 pounds per square inch). An additional factor affecting the materials yield

stress was the straightening process used during manufacturing. This process will work

6

harden the outside diameter of the nozzle resulting in the outside diameter yield stress

being substantially above inside diameter yield stress.

3.2

Probable Cause for Vessel Head Wastage Cavities

Corrosion experiments (discussed in Section 3.2.1.2) simulating a cracked nozzle have

confirmed that corrosion rates in excess of 2 inches per year are possible in low alloy

steel. Nozzle 3 contained two through-wall axial cracks, which traversed the J-groove

weld. The longest of these two cracks extended for approximately 1.3 inches above the

J-groove weld. This crack would likely be the oldest crack in this nozzle as discussed in

Section 3.1.1. The crack was on the downhill side of Nozzle 3 in direct alignment with

the long dimension of the cavity. Therefore, the AIT concluded that the cavity observed

on Nozzle 3 was associated with boric acid corrosion from crack induced leakage at this

nozzle. Further, the AIT concluded, based on corrosion products observed on the head

and in the containment air coolers and radiation element filters, that the corrosion

process had been in progress for at least 4 years.

For Nozzle 2, the crack with the longest dimension above the J-weld was also located in

the same area as the observed area of metal loss behind this nozzle. Again, the AIT

considered that the metal loss was caused by boric acid corrosion from crack induced

leakage at this nozzle.

3.2.1

Boric Acid Corrosion Mechanism

Pressurized water reactors use boric acid in the reactor coolant as one means of

controlling the nuclear reaction rate. The levels of boric acid in the reactor coolant can

range up to 2000 parts per million, which is generally not corrosive to materials used in

the reactor plant. However, if boric acid is allowed to reach a concentrated solution it

can become very corrosive to carbon steel components. The NRC issued GL 88-05,

Boric Acid Corrosion of Carbon Steel Reactor Pressure Boundary Components in PWR

[Pressurized Water Reactor] Plants, in March of 1988. The Generic Letter was in

response to several industry incidents where concentrated boric acid solution, formed by

evaporation of water from leaking reactor coolant, corroded reactor coolant pressure

boundary components. The Generic Letter requested that licensees implement a

program consisting of systematic measures to ensure that the reactor coolant pressure

boundary would have an extremely low probability of abnormal leakage, rapidly

propagating failure, or gross rupture.

3.2.1.1 Boric Acid Corrosion Processes

Compounds of boron can develop from the precipitation of boric acid from solution.

Boric acid (H3BO3) and boric oxide (B2O3) can exist in a solid or molten state. The solid

form of boric acid produced during evaporation depends on the rate of evaporation with

faster evaporation creating smaller particles. When a boric acid solution comes in

contact with boric acid crystals, larger crystals tend to form. It is also possible to form a

salt tree when previously precipitated solids form a porous structure that can wick more

solution to the vapor phase interface.

7

Boric acid solution that leaks onto the vessel head will cause the water to flash to steam,

leaving behind white, popcorn-like boric acid crystals. This form of boric acid crystals is

relatively easy to remove after the reactor is cooled down to ambient temperature. Dry,

white, powdery boric acid crystals on the reactor vessel head have been found to be

relatively benign while the reactor head is at operating temperatures. Although some

darkening of the boric acid crystals may occur with age, brown or rust colored boric acid

is a strong indication that corrosion has occurred and a problem potentially exists.

Above 302oF, boric acid begins to dehydrate to form boric oxide:

2 H3BO3  B2O3 + 3H2O

The final condition of the mixture of boric acid and boric oxide is site specific, depending

on the relative quantities of each component and the amount of flow of boric acid, the

porosity created by steam escaping, and the presence of impurities such as iron oxide.

Boric oxide begins to soften at 617oF and becomes highly viscous at 842oF.

As boric acid that is not converted to the oxide is heated above 365oF, it may become

a viscous fluid (A. S. Myerson, Handbook of Industrial Crystallization, Butterworth-

Heinemann, Boston, 1993), conforming to the surrounding geometry under the influence

of gravity. Molten boric acid can contain between 8 and 14 percent water and can be

highly corrosive under some conditions (U. Gurbuz Beker and N. Bulutcu, A New

Process to Produce Granular Boric Oxide by High Temperature Dehydration of Boric

Acid in a Fluidized Bed, Transactions of the Institute of Chemical Engineers, 74A, 133,

1996). Discussions with the NRC staff and staff members at the Brookhaven National

Laboratory indicate that the boric acid/boric oxide mixture can vitrify if concentrated

sufficiently and held at a high enough temperature.

3.2.1.2 Industry Accepted Boric Acid Corrosion Rates

In GL 88-05 corrosion rates were identified for pressure boundary materials of up to

0.019 inches per year (in/yr) at 500oF. For lower temperatures, corrosion rates up to

4.8 in/yr were identified. However, these corrosion rates were established for

configurations which were not representative of the CRDM nozzle to head annulus gap

configuration.

A Babcock & Wilcox (B&W) owners group report, BAW-10190P, Safety Evaluation For

B&W Design Reactor Vessel Head Control Rod Drive Mechanism Nozzle Cracking,

was completed in May of 1993. In this report, a Combustion Engineering pressurizer

heater sleeve mockup was used as the basis for establishing a 1.07 cubic inches per

year corrosion rate as the applicable rate for the vessel head due to cracks in CRDM

nozzles. The test results used by B&W were documented in EPRI report TR-102748S,

Boric Acid Corrosion Guidebook. This B&W analysis concluded that with this

corrosion rate, a plant would remain within ASME Code structural requirements for a

minimum of 6 years. The AIT identified a test note in the EPRI report which stated that

the maximum volume loss of 1.07 cubic inches per year may not be conservative for all

cases since the volume loss is likely to increase as the corrosion depth and wetted

surface area increase.

8

In November of 2001, an EPRI test was documented in Revision 1 to the Boric Acid

Corrosion Guidebook. This test was performed utilizing a configuration, temperature,

materials and leak rates which more closely matched the CRDM nozzle to vessel

configuration. This test identified a corrosion rate of up to 2.37 in/yr. This test also

indicated that the maximum corrosion occurred at the location where the boric acid

entered the annulus gap. The contour of the degradation observed at Nozzle 2 and

Nozzle 3 appeared to support this test result.

3.2.2

Licensee Preliminary Identified Cause

The preliminary conclusions of the licensees root cause team were documented in a

memorandum to the Davis-Besse Site Vice President, dated March 22, 2002

(Attachment C). In this memorandum, the root cause team concluded: The factors

that caused corrosion of the reactor pressure vessel (RPV) head in the regions of

nozzles #2 and #3 are the CRDM nozzle leakage associated with through-wall cracking,

followed by boric acid corrosion of the RPV low-alloy steel. The root cause team

concluded that the cracking initiated in Nozzle 3 in 1990 (+/- 3 years) and the crack had

propagated through-wall between 1994 and 1996. The average rate of RPV head

corrosion was identified as 2 inches per year along the line from Nozzle 3 to Nozzle 11.

In this memorandum, the root cause team also stated that: The estimated corrosion

rates are compatible with test results reported in Electric Power Research Institutes

(EPRI) Boric Acid Corrosion Guidebook. They are also consistent with the video,

photographic and supporting plant data, that show that significant corrosion was

occurring by the 1998 to 1999 time-frame. In addition, the root cause team identified a

number of causal factors such as boric acid accumulation on the top of the RPV head

and flange leakage.

The AIT concluded that the licensees root cause team had reviewed the applicable

historical data and established an appropriate time-line that supported the root cause.

Although the AIT agreed with the preliminary root cause conclusions, there were several

crucial questions left unanswered. The licensees root cause efforts were continuing at

the conclusion of the NRCs inspection. After the conclusion of the AIT, the licensee

provided their final root cause analysis report to the NRC, on April 18, 2002, and

provided responses to the NRCs questions associated with the preliminary root cause

report on April 30, 2002. These documents are currently under review.

4.0

HISTORY OF VESSEL HEAD INSPECTIONS AND MATERIAL CONDITION

4.1

Background CRDM Flange Leakage

Historically, CRDM flange leakage had been observed at several B&W designed plants.

At Davis-Besse, CRDM flange leakage typically resulted in deposits of boric acid on the

service structure above the reflective insulation. However, flange leakage in liquid form

also ran down the nozzles through the clearance gaps in the insulation and became

boric acid deposits on the vessel head. The access for removing the boric acid deposits

and inspecting the vessel head for corrosion is through (18) 5-inch by 7-inch rectangular

openings or weep holes. These openings are at the bottom of the service structure

9

where it is attached to the vessel head. This location combined with the curvature of the

vessel head made it difficult to inspect and clean the top center portion of the vessel

head. Visual inspections of the vessel head have typically been accomplished using

small video cameras inserted through the weep holes. Refer to Slide 5 in Attachment B

for a diagram of the vessel head.

The CRDM flanges and flange bolts are made of stainless steel, corrosion resistant

materials. Although the split nut-rings, located on the underside of the lower flange

face, are made of a low alloy steel and are susceptible to corrosion, they have been

coated with a corrosion resistant product. The nut-rings have not been found with boric

acid corrosion at Davis-Besse. Because of these corrosion resistant materials, leakage

from CRDM flanges typically does not result in corrosion, and any boric acid deposits

from flange leakage are normally white or light in color. Conversely, as documented in

the Davis-Besse Boric Acid Corrosion Control Procedure, boric acid deposits with red or

rust color indicate that corrosion has occurred.

The licensee systematically resolved CRDM flange leakage by replacing the flange

gaskets with a new design. Starting in 6 RFO (1990), gaskets were replaced on flanges

which had developed leaks during the previous operating cycle, such that by 10 RFO

(1996), the last nine old-design gaskets were replaced even though these flanges were

not leaking.

4.2

History of Flange Leakage and Reactor Head Inspections

Inspections of the reactor head associated with identifying boric acid deposits were

recorded after the licensee established a Boric Acid Control Program in 1988 in

response to NRC GL 88-05. The following inspection results were documented in the

licensees corrective action system through PCAQRs [potential conditions adverse to

quality reports] or CRs [condition reports] and/or recorded on video-tapes:

In April of 1990 (6 RFO) 22 leaking CRDM flanges were identified and repaired

(PCAQR 90-0120).

In September of 1991 (7 RFO) 15 out of 21 leaking CRDM flanges were

repaired. Boric acid was observed on the reactor vessel head that ran along the

curvature of the head and stopped on the vessel closure bolts (PCAQR 91-

0353). The source of these deposits was identified as flange leakage. Cleaning

was performed with a wire brush and vacuum. No surface irregularities were

noted following cleaning; however, the extent of deposits if any that remained

after cleaning was not documented.

In March of 1993 (8 RFO) 14 leaking CRDM flanges were identified and 11 were

repaired (PCAQR 93-0132). The boric acid from flange leakage was removed to

the extent possible by washdown of the head (PCAQR 96-551). The AIT viewed

a videotape of the head inspection conducted during this outage and prior to the

head washdown. Discrete patches of brown and white boric acid deposits were

observed which were more numerous near the center of the head.

10

In October of 1994 (9 RFO) eight CRDM flanges were leaking. All eight were

repaired including three leaking flanges from the previous outage (PCAQR 94-

0912). No record of a reactor vessel head inspection could be found.

In April of 1996 (10 RFO) the remaining nine CRDM flanges (non-leaking) not

previously repaired were modified with an enhanced gasket design. The

head was inspected and video-taped using a remote camera mounted to a

hand-held pole inserted through the weep holes. Several patches of boric

acid accumulation were identified including a brown stained deposit at

Nozzle 67 (PCAQR 96-551). The licensee documented that boron deposits

could be indicative of flange leakage or nozzle leakage. A vacuum was used to

remove boric acid deposits, but was not fully effective at removing the deposits

of boric acid near the center of the head. The corrosion on the head from

remaining boric acid was evaluated and considered minimal based on B&W

Document 51-1229638, which identified minimal boric acid corrosion of carbon

steel head material at temperatures corresponding to the normal head operating

temperature. The licensee concluded that 50 to 60 percent of the head had

been examined during this inspection. The limited head examination appeared

to be due to access restrictions caused by the weep hole access limitations and

the curvature of the head. The AIT observed the videotaped inspection and

noted that the boric acid deposits were generally white in color and appeared to

be the consistency of loose powder and discrete lumps.

In May of 1998 (11 RFO) one leaking CRDM flange was identified and not

repaired (PCAQR 98-0649). The head was inspected and video-taped using a

remote camera mounted to a hand-held pole inserted through the weep holes.

This inspection identified areas near the center of the head covered with an

uneven layer of boric acid (PCAQR 98-767). The licensee documented that the

boric acid deposits were removed as best as we can. The boric acid color was

rust brown, which the licensee attributed to old deposits of boric acid. The

previous root cause investigation and source documents from PCAQR 96-551

were referenced as the basis for leaving boric acid deposits on the head. The

licensee concluded that due to the minimal operating time below 550F, there

was no impact on vessel head integrity. Based on review of this video-taped

inspection, the AIT identified consolidated boric acid deposits near the center

region including Nozzle 2 and 3 locations. On the head at an elevation below

Nozzles 3 and 11, the AIT noted that the boric acid appeared highly adherent

and rust brown in color.

In April of 2000 (12 RFO) five leaking CRDM flanges were identified and repaired

(CR 2000-0782). The head was inspected and video-taped using a remote

camera mounted to a hand-held pole inserted through the weep holes. Lava-

like brown/red deposits of boric acid over 1-inch thick were observed on much of

the vessel head (CR 2000-1037). The corrective action for this condition was to

repeat cleaning of the head until most of the boric acid deposits are removed.

Licensee logs recorded that crowbars were needed to remove the solid rock

hard deposits of boron on the head. In addition, pressurized heated water was

used to remove the boric acid deposits. The extent of remaining boric acid

deposits or evaluation of the effects on the head was not documented in the

11

corrective action system after this cleaning. The system engineer also reported

a large amount of boric acid deposits were observed above the mirror insulation

due to flange leakage. The AIT viewed the video-taped examination made with a

remote camera after the cleaning. This videotape showed a thick layer of lava-

like brown/red boric acid that remained around the nozzles in the center of the

head.

In February of 2002 (13 RFO) no CRDM flange leakage was identified.

The head was inspected and video-taped using a remote camera mounted

to a hand-held pole inserted through the weep holes. The licensee

documented that more boron than expected was found on the top of the head

(CR 02-00685). Because the head was covered with boric acid and debris

deposits, indications of nozzle crack induced leakage could not be positively

identified at any nozzle location. The AIT reviewed pictures and tapes of this

head inspection, which showed a thick lava-like brown/red deposit of boric acid

covering the center of the head. Specifically, for 12 nozzles near the center of

the head, the boric acid layer was several inches thick and precluded access for

the remote camera inspection. The licensee subsequently removed the boric

acid deposits from the head using hot pressurized water and identified the large

head cavity at Nozzle 3.

The AIT noted the following important aspects in the above history of inspections and

material condition of the RPV head:

(1)

No flange leakage was found during 10 RFO (1996), and very limited flange

leakage was noted during 11 RFO (1998). However, boric acid accumulation on

the reactor vessel head increased from 9 RFO (1994) to 10 RFO (1996) and

from 10 RFO (1996) to 11RFO (1998). Although the boric acid accumulation did

not come from flange leakage, the licensee apparently did not deduce that it then

must have come from pressure boundary leakage, such as nozzle cracking.

(2)

Although five flanges were documented as leaking during 12 RFO (2000),

according to CR-2000-0782, only four of the flanges showed positive evidence of

gasket leakage. The fifth flange did not show the typical signs of flange leakage,

but boric acid deposits had built up under the flange to the extent that the flange

could not be fully inspected. This flange was for Nozzle 3, and the licensee

concluded that the boric acid buildup was due to the flange leaking. The

licensee apparently did not consider that the boric acid buildup could be due to

nozzle leakage from below.

(3)

Pictures of the reactor vessel, attached to CR-2000-0782, showed rust colored

boric acid deposits emanating from the inspection openings on the reactor

vessel head service structure. Although the licensees boric acid corrosion

control procedure specifically stated that corrosion will most likely be exhibited by

rust stained boric acid, the source of these corrosion products was not

addressed in the condition report.

12

5.0

OPPORTUNITIES FOR EARLY DETECTION OF HEAD DEGRADATION

The AIT evaluated plant indications that could have provided an early opportunity to

detect the corrosion occurring in the vessel head. The AIT identified the following

indicators which could have provided early detection of the head corrosion.

5.1

Boric Acid Corrosion Control Program

Leakage from the reactor coolant system (RCS) with the reactor at power will flash to

steam and leave behind boric acid crystals. Averaged over the course of a fuel cycle,

there is approximately 0.03 pounds of boric acid per gallon of primary coolant.

Assuming a leak rate of 0.001 gallons per minute, approximately 15 pounds of boric acid

crystals would be produced in the vicinity of the vessel head by a postulated crack in a

CRDM nozzle over one year. This leak rate would be significantly less than the

minimum detection capability of the plant leakage detection systems. Therefore,

inspection of the reactor head for boric acid deposits is potentially the most sensitive

method available for detecting small leaks caused by cracked nozzles. However, there

are limitations to this method. First, depending on location, a leak may not be

accessible with the reactor at power. Consequently, certain leaks can only be identified

when the reactor is shut down, which may only occur during refueling outages every two

years. Second, this method depends on removing all existing boric acid accumulation,

so any new leak can be detected without being masked by previous accumulations.

This is critical because very small leaks may not be identifiable if the preexisting

accumulation is not removed.

As previously discussed in Section 4.2, the licensee had preformed visual inspections

of the reactor vessel head in 7 RFO (1991), and 8 RFO (1993) in accordance with

GL 88-05 guidance. Davis-Besses implementing procedure for GL 88-05 was

NG-EN-00324, Boric Acid Corrosion Control. Although recurring CRDM flange

leakage was documented during 9 RFO (1994), licensee personnel were unable to

identify any records documenting the visual inspections of the head during that outage.

In addition, boric acid deposits have historically been left on the head from flange

leakage as discussed in Section 4. A leaking flange typically results in boric acid

deposits which travel down past the head insulation resulting in a deposit/buildup of

boric acid on the head. In accordance with the boric acid control program, these

deposits should have been removed and the head inspected and any corrosion

evaluated.

During 10 RFO (April 1996), a licensee engineer initiated PCAQR 96-0551,

Boric Acid on Reactor Vessel Head, to document that the steps required by

Procedure NG-EN-00324, Boric Acid Corrosion Control, had not been followed during

the previous outage and that the procedure could not be fully implemented due to

limited access to the reactor vessel head. The evaluation presented in this PCAQR

acknowledged the need to clean the vessel head, such that nozzle leakage could be

detected in the future. Also, the initial assessment in this PCAQR stated that the failure

to clean the boric acid deposits made it difficult to determine if the deposits occurred

13

because of leaking flanges or because of a crack in the CRDM nozzle. Licensee

managers approved the PCAQRs initial assessment subject to the following comment:

Nozzle cracking is of course a significant issue. However, at present, the

probability of occurrence is relatively low. We should remove boron from the

reactor pressure vessel head as best we can and so as to minimize dose. This

will allow us to monitor any leakage, should a nozzle crack initiate.

The corrective action for this PCAQR became a Request for Modification 94-0025 (see

Section 5.5.1 for additional discussion on the delay of this modification).

Because of access limitations (see Sections 4.1 and 4.2), the RPV head was not

completely cleaned and some portions were not thoroughly inspected, as specified by

the licensees Boric Acid Corrosion Control Program. The bases for not cleaning or

inspecting the CRDM nozzles near the center of the RPV head was documented in

PCAQRs or provided by licensee staff during interviews with the AIT. Specifically, the

following information was utilized by the licensee to justify leaving boric acid deposits on

the RPV head as identified during inspections in 10 RFO, 11 RFO and 12 RFO:

1) B&W Owners Group stress analyses had predicted that peripheral nozzles

were more likely to crack than nozzles near the center of the vessel head.

2) Dried boric acid was not corrosive to the vessel and moderate amounts of

boric acid from CRDM flange leakage had historically been found and

cleaned up in the past, with no vessel corrosion.

3) Very limited boric acid corrosion occurs in the temperature range existing at

the vessel head.

4) EPRIs Boric Acid Corrosion Guidebook indicated that, under specific

circumstances, a layer of boric acid potentially protects a surface from

ongoing corrosion by keeping water away from the surface.

5) CRDM nozzle cracking was an age related phenomenon, and the

Davis-Besse staff believed they should not see any cracking because it was

several years younger than Oconee where significant problems had not yet

occurred. This was codified by the B&W Owners Group in July 1997 through

a probabilistic susceptibility ranking that was developed in response to the

NRCs GL 97-01.

The identification of nozzle cracks at Oconee Units 1 & 3, prompted the NRC to issue

Bulletin 2001-01, which requested licensees to provide information, including a

description of their previous inspections of the reactor vessel head. The Davis-Besse

responses of September 4 and October 17, 2001, described their previous inspection

and noted that, since 1996, four of the nozzles in the center of the vessel head were

obscured with boric acid deposits and could not be viewed. In addition, the licensees

responses described their analytical efforts to verify that gaps would exist between the

CRDM nozzles and the reactor vessel head, permitting through-wall leakage from a

crack in a nozzle to be observed via boric acid deposits.

The licensees analyses concluded that, except for Nozzles 1, 2, 3, and 4 (center

nozzles), gaps would exist during normal operating conditions through which leakage

could occur and boric acid deposits would be evident. In their supplemental response to

14

the NRC Bulletin, dated October 30, 2001, the licensee stated that based on the above

analytical results, the Davis-Besse staff would not expect to see boric acid residue

around Nozzles 1, 2, 3, or 4 if a crack were present. This was based on the

manufactured interference fit between the nozzles and the vessel head. The notable

aspect of this conclusion was that the analytically predicted interferences ranged from

0.000025 to 0.000004 inches. Because the fabrication tolerances were more than an

order of magnitude greater than the analytical results, the AIT considered the licensees

conclusion, relative to not expecting boric acid residue if a crack were present in these

nozzles, to be unrealistic.

During interviews with the AIT, licensee personnel acknowledged that the reactor vessel

head was treated less rigorously than other components in the plant, within the context

of the GL 88-05 program. Although the boric acid corrosion control program was

appropriately entered when boric acid was identified on the reactor vessel head, the

resolution of the issue was not treated the same. Using the longstanding rationale

discussed above, the licensee used a philosophy that boric acid had been on the reactor

head for many years and no problems had ever been found.

5.2

Reactor Coolant System Leakage Detection

Because leakage from the through-wall cracks in Nozzle 3 would result in reactor

coolant leakage into the containment atmosphere, the leakage detection systems in

containment were reviewed to determine whether this system could have provided an

early indicator of head corrosion. The observed leakage rate from a cracked nozzle

would be expected to be very small based on a leakage rate (0.003 gallons per minute

(gpm)) attributed to CRDM nozzle cracks observed at a foreign reactor plant (Bugey).

Regulatory Guide 1.45, Reactor Coolant Pressure Boundary Leakage Detection

Systems, details requirements for leakage monitoring equipment such as the

containment atmosphere particulate and gaseous radioactivity monitoring systems and

containment sump level/flow monitoring system. The licensee has implemented a leak

detection program in accordance with Regulatory Guide 1.45 as described in the

Updated Safety Analysis Report, Section 5.2.4.

Reactor coolant system (RCS) leakage is grouped into two categories: identified and

unidentified. Identified leakage is that which is captured and metered through closed

systems, such as a collecting tank (e.g., pump seals and valve packing leaks); leakage

into containment atmosphere from sources that are both specifically located and known

not to interfere with the operation of leakage detection systems or not to be pressure

boundary leakage; leakage through the steam generators to the secondary system; and

reactor coolant pump seal returns. Unidentified leakage is everything which is not

identified leakage.

Unidentified RCS leakage was normally less than 0.1 gpm (monthly average), until

October of 1998, when a decision was made to remove the rupture disks downstream of

the pressurizer relief valves for design concerns (PCAQR 98-1980). Specifically, a drain

line, designed to collect relief valve leakage in the quench tank, was bypassed in this

modification. This allowed leakage past the relief valves to be vented directly into the

containment atmosphere, which collected in the normal sump and added to the

15

unidentified leakage, which increased to a maximum of 0.8 gpm. During a mid-cycle

outage in May of 1999, the licensee resolved this design concern by installing new

rupture disks and reconnecting the drain line. This resulted in a decrease in

unidentified leakage. However, the unidentified leakage returned to levels between

0.15 and 0.25 gpm. Subsequent investigations and containment entries were not

successful in identifying definitive sources of this leakage. The licensee concluded,

based upon the history of CRDM flange leakage and that unidentified leakage values

observed at Davis-Besse were near industry averages, the leakage was most likely from

the CRDM flanges.

Because of historical variations in unidentified leakage compared to the relatively small

amount of leakage associated with CRDM cracks, the AIT concluded that, by itself,

unidentified leakage trends were not a reasonable method of detecting nozzle cracking.

However, when considered together with other indications of corrosion products as

discussed in Section 4.2, above, and in Sections 5.3 and 5.4 below, the AIT concluded

that this was a missed opportunity to detect the corrosion occurring on the reactor

vessel head.

5.3

Containment Air Coolers

Reactor coolant leakage through the cracks in Nozzle 3 would travel as steam and liquid

in the annulus behind the nozzle and leave boric acid deposits on the top of the head.

In addition, this steam leakage would cause boric acid and corrosion products from the

head cavity to be divided into fine particles which would be dispersed into the air space

above the head. These fine particles would then be captured by the service structure

ventilation system intake and be distributed throughout the containment. A key area

which could collect these airborne particles of boric acid and corrosion products is at the

containment air coolers (CAC).

The vessel head service structure ventilation pulls a suction from the CRDM flange

area through the fans located on the 603 feet elevation, exhausting through ductwork

to the top of the East D-ring. This provided a potential pathway for any corrosion fines

and boric acid particulate dispersion originating from the vessel head. In November of

2001, radiological surveys showed a contamination plume effect originating from the

service structure ventilation exhaust over the East D-ring. However, an isotopic analysis

was not performed of the plume to fully characterize the source of the contamination.

Additionally, two containment recirculation fans provide a mixing of the containment

atmosphere, further dispersing the fines and particulates.

The CAC system consists of three separate tube/fin coolers (which are cooled by the

service water (SW) system) located inside containment, and connected to a common

supply plenum. Downstream of this plenum is a ductwork distribution system, designed

to distribute air over and around all heat producing equipment, such as the reactor

vessel, D-rings (housing the steam generators, pressurizer and reactor coolant pumps)

and incore instrument tank. The external surfaces of the cooler tube banks are readily

visible from the outside of the coolers, and have a remote indication of plenum pressure

(used to determine cooling fin fouling) in the control room.

16

If a leak occurs from the RCS during normal operations, an aerosol mist is produced

from the water flashing and evaporating as it exits the leak, increasing containment

ambient humidity. Since the inlet water temperature of SW to the CACs is normally

between 40°F and 75°F, substantially cooler than containment air temperatures, the

CACs condense this ambient humidity to water, which is ultimately collected in the

normal containment sump. In the process of removing the humidity, the CACs also

collect particulate boric acid (which would be released with the RCS leakage as fine

particles) on the cooling fins, in the discharge plenum and the associated ductwork.

This fouling will decrease the plenum pressure, as read remotely in the control room,

during periods of high boric acid accumulation.

In 1992, the licensee had experienced a CAC fouling from a leak in the reactor head

vent line flange to the primary side of the steam generator. As a result, the licensee

cleaned the boric acid, evident by the uniformly white coating on all three coolers. After

repairs to the flange, no further boric-acid precipitated cleanings were required for

several years.

In October of 1998, the removal of the rupture disks downstream of the pressurizer relief

valves substantially contributed to the RCS unidentified leakage. In November 1998,

PCAQR 98-1980 identified that the CAC fouling had increased correspondingly to

increased leakage from the pressurizer reliefs. The CACs were cleaned 17 times from

November 1998 to May 1999. During a mid-cycle outage in May 1999, the design

concern was resolved, the rupture disks reinstalled, and the drain line reconnected.

However, two additional CAC cleanings were conducted, one in June 1999 and one

in July 1999. The post-job critique observed the boric acid to be rust color on and

in the boron being cleaned away from CAC No. 1. Subsequent interviews indicated

this was presumed to be the result of restoring from the mid-cycle outage, and the

residual humidity in containment from outage-related repairs. After being cleaned in

July 1999, the CACs did not need any further cleaning for approximately 10 months.

Although the licensee installed high efficiency particulate air filters (inside containment)

during August and September 1999, this did not appear to factor into the need for CAC

cleaning.

After 12 RFO (May of 2000), CAC deposits were again forming, as evidenced by the

decrease in plenum pressure. Eight CAC cleanings were conducted between

June 2000 and May 2001, with no further cleanings required through the end of cycle.

However, for 13 RFO (February 2002), the licensee reported (15) 5-gallon buckets of

boric acid were removed from the ductwork and plenum. Significant boric acid was

found elsewhere within containment, including on SW piping, stairwells and other areas

of low ventilation.

After the 1999 mid-cycle outage, the licensee had attributed the excessive boric acid

accumulation and CAC cleanings to leakage from CRDM flanges. In 12 RFO

(May 2000), several leaking flanges were repaired, the results of which could not be

verified throughout the cycle. However, 13 RFO (February 2002) inspections indicated

the repairs had been successful, and no flange leakage was detected. Furthermore,

earlier experience with leaking flanges (pre-1992, and 1992-1998) did not result in the

need to clean the CACs. Therefore, CRDM flange leakage would not have reasonably

been the major contributor to the increased boric acid loading on the CACs during this

17

time frame. The licensee had also attributed the discoloration of the boric acid to

migration of the surface corrosion on the CACs into the boric acid and the aging of the

boric acid itself.

The AIT considered the sudden change to rust colored boric acid deposits in June of

1999, to indicate corrosion product accumulation from the formation of the head cavity

near Nozzle 3. The failure of the licensee to identify the source of these deposits

represented a missed opportunity to identify the corrosion cavity in the head at that time.

5.4

Radiation Elements

As discussed in Section 5.2, steam leakage through the cracks in Nozzle 3 would result

in fine particles of boric acid and corrosion products. These particles would then be

captured by the service structure ventilation system intake and distributed throughout

the containment. An area where these fine particles of boric acid and corrosion

products would be collected and observed is in the radiation element (RE) system filters.

There are two identical radiation element air sampling systems, drawing from two

sample locations within containment. Air samples are drawn from within containment,

passed through a particulate filter, an iodine sample cartridge and a noble gas detector

before being exhausted back into containment. Both systems normally draw a sample

from near the top of the D-ring structures, but can also draw from near the polar crane,

and near the personnel airlock on the 603 feet elevation.

Boric acid accumulation on the RE filters can clog the filters and decrease flow to below

acceptable levels, necessitating a filter change. Licensee records correlate past RCS

leakage increases with RE filter changes, such as in 1992 when the reactor head vent

flange leakage caused this to occur. In March of 1999, RE filter clogging from boric acid

deposits was attributed to the pressurizer relief valve rupture disk maintenance which

occurred in 1998. Filter changes normally occurred based on a monthly schedule rather

than low flow rates. Beginning in May of 1999, the schedule of filter change out went

from a monthly interval to an irregular 1 to 3 week interval, occasionally dropping to a 1

to 2 day interval by November 1999. In response to the increased frequency of filter

changeouts, the licensee installed two large high efficiency particulate air filter units

inside containment to capture a large portion of the corrosion fines. Additionally, the RE

sample points were changed to the alternate locations. This action appeared to improve

the service life of the filters, but did not eliminate the filter loading conditions completely.

In May of 1999, the RE filters began accumulating a yellowish-brown material. This

material was sent to an external laboratory for analysis. The results of this analysis

were received in November 1999, and positively identified the presence of ferric oxide.

Specifically, this analysis stated, The fineness of the iron oxide (assumed to be ferric

oxide) particulate would indicate it probably was formed from a very small steam leak.

The particulate was likely originally ferrous hydroxide in small condensed droplets of

steam and was oxidized to ferric oxide in the air before it settled on the filters; and the

iron oxide does not appear to be coming from the general corrosion of a bare metal

surface in containment or from steam impingement on a metal surface.

18

Accumulation of boric acid on the RE filters was readily recognized as a symptom of

RCS leakage. During 12 RFO, CRDM flange D10 was attributed as the source of the

RCS leakage, since the flange required machining to correct the leakage. However, the

presence of ferric oxide fines was not explained, nor were multiple containment entries

successful in determining a source. Additionally, past CRDM flange leakage had not

significantly contributed to the CAC fouling, nor the RCS leakage indications.

Therefore, the AIT believed that the corrosion deposits first identified in the RE filters

beginning in May of 1999, indicated that corrosion was occurring due to the formation of

the head cavity near Nozzle 3. The failure of the licensee to identify the source of these

corrosion products represented a missed opportunity to identify the corrosion cavity in

the head at that time.

5.5

Causal Factors Influencing Head Degradation Detection

Several decisions made by Davis-Besse personnel at various times directly influenced

or potentially affected their ability to detect the head degradation associated with the

CRDM nozzle leakage. These are discussed below.

5.5.1

Decision to Delay Modification to Service Structure

In March of 1990, modification 90-0012 was initiated to install multiple access ports in

the service structure to permit inspection and cleaning of the vessel head. This

modification was canceled in 1992, because the current inspection techniques were

considered adequate.

In March of 1994, a licensee engineer initiated PCAQR 94-0295 to question why there

was no commitment requiring a visual inspection of the reactor vessel head every

refueling outage, as referenced in the NRC 1993 Safety Evaluation for the Alloy 600

CRDM nozzle cracking issue. The PCAQRs response from the Nuclear Assurance

Director indicated that the commitment for the visual inspection did not appear to have

been a licensee commitment to the NRC. Regulatory Affairs and Design Engineering

personnel indicated that, although an enhanced visual was not a commitment to the

NRC, they recommended the visual inspection be done. However, the plant engineering

staffs comment in the PCAQR stated that there was a low risk of a crack in CRDM

nozzles since none had been identified in the United States, and that the available

inspection methods were not highly reliable. On that basis, the plant engineer felt it was

not necessary to perform the inspections.

In May of 1994, the licensee engineer who wrote the above PCAQR initiated a Request

for Modification (RFM 94-0025) to install openings in the CRDM service structure to

allow thorough inspection and cleaning of the reactor vessel head. The modification

request noted that, out of all of the B&W plants, only Davis-Besse and Arkansas

Nuclear One, Unit 1, had not installed the access openings in the service structure. The

modification request cited the following reasons for the modification:

1) there was no access to the reactor vessel head or CRDM nozzles without

the modification, and there was an ongoing industry concern for Alloy 600

nozzle cracking;

19

2) inspection of the reactor vessel head for boric acid corrosion was difficult

and not always adequate, because the video inspections did not encompass

a 100 percent inspection of the head;

3) cleaning boric acid residue from the vessel head did not encompass

100 percent, because the size and geometry of the weep holes only

permitted cleaning of the lower one-third of the head with scrapers and wire

brushes.

The modification was approved by the plant in July of 1994, but remained unfunded by

the Project Review Committee/Project Review Group until November of 1998, when it

was scheduled for implementation in 13 RFO (2002). The modification was

subsequently deferred until 14 RFO by the Project Review Group, as part of an effort

to meet the 2001/2002 expenditure targets by reducing the number of projects

implemented. In discussing the reasons for not implementing this modification, the

rationale identified in Section 5.1 were also applied. The AIT considered the delay in

implementing the modification as contributing to the failure to detect head degradation.

5.5.2

Decision to Delay Repair of CRDM Flange on Nozzle 31 in 11RFO

During 8 RFO (1993), CRDM flange leakage was noted on several CRDM flanges

including the flange for Nozzle 31. The corrective actions included polishing the flange

surface and replacing the gasket with a new design. The PCAQR issued to document

this condition (93-0132) contained a recommendation that the flange surface be

inspected during each subsequent maintenance outage and be machined if further

leakage occurs. During 11 RFO (1998), the CRDM flange for Nozzle 31 was found to

be leaking, and as indicated in PCAQR 98-0649, the amount of leakage was not

considered significant compared to flange leakage from previous outages.

Consequently, no corrective actions were taken, even though the vendor (Framatome)

reiterated their recommendation from 1993 to machine the flange. The PCAQR did

contain a recommendation to reexamine the flange for Nozzle 31 during 12 RFO and to

replace the gasket if the flange was leaking.

During 12 RFO (2000), significant flange leakage was noted and five leaking flanges

were identified during the video inspections of the CRDM flanges, including Nozzle 31s.

The majority of the boric acid accumulation was attributed to Nozzle 31s flange due to

steam cutting of the flange face. Condition Report 2000-1037 was written to describe

the boric acid accumulation on the RPV head and on top of the insulation. The boric

acid accumulation was attributed to leaking CRD flanges. The AIT considered the delay

in repairing Nozzle 31s flange as a contributing cause of this event, because the

extensive amount of flange leakage contributed to the boric acid deposits on the head

which masked evidence of the nozzle leakage occurring at this time.

6.0

CONCLUSIONS

The AIT presented the inspection results to Mr. Saunders and other members of the

licensee management at the conclusion of the inspection on April 5, 2002. The licensee

acknowledged the conclusions presented as discussed in Attachment B and

summarized below.

20

The AIT concluded that the probable cause of the cavity at Nozzle 3 was boric acid

corrosion of the head associated with reactor coolant leakage from a through-wall crack

in this nozzle. Further, the AIT concluded based on corrosion products observed on the

head, and in the CAC and RE filters that the corrosion process had been in progress for

at least 4 years.

The AIT concluded that the probable cause of the cracking observed in the five

penetration nozzles was PWSCC. This was based on similar cracking identified at two

other B&W plants that performed destructive analysis of cracked nozzles fabricated

from the same heat of material to confirm PWSCC.

The AIT evaluated the indications which existed that could have provided an early

opportunity to detect evidence of the formation of the corrosion cavity in the head at

Nozzle 3. The AIT identified several opportunities which were available to the licensee

to potentially identify this corrosion cavity at an earlier point in time. Specifically, these

missed opportunities were associated with the failure to identify the source of the

corrosion products deposited in the CAC and RE filters in early 1999 and the failure to

remove boric acid or evaluate the source of corrosion products which accumulated on

the vessel head.

21

KEY POINTS OF CONTACT

DAVIS-BESSE

H. Bergendahl, Vice President - Nuclear

D. Eshelman, Director, Support Services

R. Fast, Plant Manager

D. Geisen, Manager, Design Engineering

D. Lockwood, Manager, Regulatory Affairs

J. Messina, Director, Work Management

D. Miller, Supervisor, Compliance

S. Moffit, Director, Technical Services

R. Saunders, President, FirstEnergy Nuclear Operating Company

NUCLEAR REGULATORY COMMISSION

J. Davis, Sr. Material Engineer, NRR

R. Gardner, Chief, Engineering Branch

J. Grobe, Director, Division of Reactor Safety

C. Lipa, Chief, Reactor Projects Branch 4

B. Sheron, Associate Director for Project Licensing and Technical Analysis, NRR

LIST OF ACRONYMS USED

AIT

Augmented Inspection Team

ASME

American Society of Mechanical Engineers

B&W

Babcock and Wilcox

CAC

Containment Air Cooler

CR

Condition Report

CRDM

Control Rod Drive Mechanism

EPRI

Electric Power Research Institute

GL

Generic Letter

gpm

Gallon Per Minute

in/yr

Inches Per Year

NRC

Nuclear Regulatory Commission

PCAQR

Potential Conditions Adverse to Quality Report

PDR

Public Document Room

PWSCC

Primary Water Stress Corrosion Cracking

RCS

Reactor Coolant System

RE

Radiation Element

RFO

Refueling Outage

RPV

Reactor Pressure Vessel

SW

Service Water

22

LIST OF DOCUMENTS REVIEWED

Calculation

SIA Calc W-ENTP-11Q-306

Finite Element Gap Analysis of CRDM Penetrations

(Davis-Besse), October 8, 2001.

Condition Reports (CR)

1992-0139

Boron Found on Containment Air Sample Filter

1993-0187

Boric Acid Accumulation on SW Piping

1998-0020

Multiple Problems Identified with RC-2

1998-0330

Industry Event (Prairie Island) Crack in the Motor Tube of the Control Rod Drives

1998-1963

Design Over-Stress of the Pressurizer Nozzles for Safety Valve

1999-0372

Received Computer PT-RE4597AA/AB High

1999-0510

Low Flow Alarm Observed on RE4597BA While Out of Service for Maintenance

1999-0745

Small Clumps of Boric Acid Present on Wall Opposite of DH108

1999-0861

RE4597AA Sample Lines Were Found to be Full of Water

1999-0928

Increased Frequency of Particulate and Charcoal Filters for RE 4597BA Being

Changed

1999-0998

Awareness of Approaching the Tech Spec Limit for Maximum Ctmt Air Temp

1999-1300

Analysis of CTMT Radiation Monitor Filters

1999-1614

Due Date of LER Commitment Missed: Boric Acid Control Program Procedure

Change

2000-0781

Leakage from CRD Structure Blocked Visual Exam of Reactor Vessel Head

Studs

2000-0782

Inspection of Reactor Flange Indicated Boric Acid Leakage From Weep Holes

2000-0903

Two of 40 CRDM Hold Down Bolts Had Indications Found During VT-1

Inspection

2000-0994

RV Head CRDM Nozzle at Location F-10 has Large Pit in Outer Gasket Groove

2000-0995

RV Head CRDM Nozzle Flange at Location D-10 has Extensive Pitting Across

the Outer Gasket Groove. Inner Gasket Also Has Pitting

2000-1037

Inspection of Reactor Head Indicated Accumulation of Boron in Area of the CRD

Nozzle Penetration

2000-1210

During Installation of Control Rod Drive Assembly at Location D-10, on the

Reactor Head, it was Discovered that Top of Motor Tube for this Drive was out of

Line with Surrounding Motor Tubes

2000-1547

CAC Plennum Pressure Drop Following 12 RFO

2000-4138

Frequency for Cleaning Boron From CAC Fins Increased to Interval of

Approximately 8 weeks

2001-0039

CAC Plenum Pressure Experienced Step Drop

2001-0487

Certain Areas Inside CTMT in Year 2000 Seeing Higher Temperatures

2001-0890

Unidentified RCS Leak Rate Varies Daily by as Much as 100 percent of the

Value

2001-1110

Chemistry is Changing Filters on RE4597BA More Frequently

2001-1822

Frequency of Filter Changes for RE4597BA is Increasing

2001-1857

RCS Unidentified Leakage at .125 to .145 gpm

2001-2012

NRC Issuance of IEB 01-01 Circumferential Cracking of RX Pressure Vessel

Head Penetration Nozzles

23

2001-2769

RE2387 Identified Spiked Above ALERT and High Setpoints

2001-2795

RE4597BA Alarmed on Saturation

2001-2862

Calculated Unidentified Leakage for Reactor Coolant System has Indicated

Increasing Trend

2001-2936

Monthly Functional Test for RE4597BA/BB Count Not Performed

2001-3025

Increase in RCS Unidentified Leakage

2001-3411

Received Equipment Fail Alarm for Detector Saturation on RE4597BA

2002-0685

Loose Boron 1-2" deep 75% Around Circumference of Flange

2002-0846

More Boron Than Expected Found on Top of Head

2002-0891

UT Performed on #3 CRDM Nozzle Revealed Indication of Through-Wall Axial

Flaws

2002-0932

Completion of UT on All 69 CRDM Nozzles Revealed Additional CRDM Cracks

Beyond #3 Nozzle

2002-1053

While Machining Reactor Vessel Head Nozzle #3 the Nozzle Machining Tool

Moved Approximately 15 Degrees

2002-1128

Evaluation of Bottom up Ultrasonic Test Data in Area of RX Pressure Vessel

Head Nozzle #3 Shows Significant Degradation of RX Vessel Head Pressure

Boundary

2002-1159

During Video Tape Review, Indication Found on Newly Machined Face on Mid-

Span of CRDM Nozzle. Appears to be Through-wall in Immediate Vicinity of

Base Metal Indications.

Drawings

M-503-127-3

Closure Head Assembly, Revision 3

M-503-212-1

Closure Head Subassembly Drawing, Revision 1

M-503-213-2

Closure Head Subassembly Drawing, Revision 2

03-1221681-03

Framatome Drawing of RV Nozzle/Nur Ring Modification

Modifications

MOD 90-0012

Modification Reactor Closure Head Access Ports

MOD 94-0025

Install Service Structure Inspection Openings

TM 1998-0036

Temporary Modification: Preliminary Evaluation of Pressurizer Nozzles for

Relief Valves Demonstrates that an Overstress Condition May Exist in the

Nozzle Flange

NRC Generic Communications for Control of Boric Acid Corrosion

IN 80-27

Degradation of Reactor Coolant Pump Studs, dated June 11, 1980

IEB 82-02

Degradation of Threaded Fasteners in the Reactor Coolant Pressure

Boundary of PWR Plants, dated June 2, 1982

IN 82-06

Failure of Steam Generator Primary Side Manway Closure Studs, dated

March 12, 1982

IN 86-108

Degradation of Reactor Coolant System Pressure Boundary Resulting from

Boric Acid Corrosion, dated December 29, 1986

IN 86-108

Supplement 1, dated April 20,1987

IN 86-108

Supplement 2, dated November 19, 1987

IN 86-108

Supplement 3, dated January 5, 1995

24

GL 88-05

Boric Acid Corrosion of Carbon Steel Reactor Pressure Boundary

Components in PWR Plants, dated March,17, 1988

IN 90-10

Primary Water Stress Corrosion Cracking (PWSCC) of Inconel 600, dated

February 23,1990

IN 94-63

Boric Acid Corrosion of Charging Pump Casing Caused by Cladding Cracks,

dated August 30, 1994

IN 96-11

Ingress of Demineralizer Resins Increases Potential for Stress Corrosion

Cracking of Control Rod Drive Mechanism Penetrations, dated

February 14,1996

GL 97-01

Degradation of CRDM/CEDM Nozzle and other Vessel Closure Head

Penetrations, dated April 1, 1997

IN 2001-05

Through-wall Circumferential Cracks of Reactor Pressure Vessel Head

Control Rod Drive Mechanism Penetration Nozzles at Oconee Nuclear

Station, Unit 3, dated April 30, 2001

Bulletin 2001-01

Circumferential Cracking of Reactor Pressure Vessel Head Penetration

Nozzles, dated August 3, 2001

Other Documents

RAS02-00132

Probable Cause Summary Report for CR2002-0891, dated March 22, 2002

NPE-96-00260

Control Rod Drive Nozzle Cracking, dated May 8, 1996

Books

RCS System Performance Books Volumes 1 though 11

BAW-10190P

Safety Evaluation For B&W Design Reactor Vessel Head Control Rod Drive

Mechanism Nozzle Cracking, dated May of 1993

BAW-10190P,

Addendum 1

B&W Owners Group Proprietary, External Circumferential Crack Growth

Analysis for B&W-Design Reactor Vessel Head Control Rod Drive

Mechanism Nozzle Cracking, dated December 1993

BAW-2301

B&W Owners Group Proprietary, B&WOG Integrated Response to Generic Letter 97-01, dated July 1997

Exam Report

Reactor Vessel Head ID Clad Thickness Measurements in Region of

Wastage Between Nozzles 3 and 11, dated March 18, 2002

Exam Report

DB-5

CRDM Nozzle 46 J-Groove Weld, dated March 24, 2002

Examination Report

Davis Besse 13 RFO CRDM Nozzle Examination Report, dated

March 11, 2002

Framatome

51-5015818-00

Davis-Besse CRDM Nozzle Heat Information, 2002

EPRI Report

TR-102748s

Boric Acid Corrosion Guidebook, Revision 0, dated April 1995

EPRI Report

1000975

Boric Acid Corrosion Guidebook, Revision 1, dated November 2001

Report 2779

Oconee Unit 3 CRDM Nozzle Crack and Material Characterization - Oconee

Unit 1 Thermocouple Tube material Characterization - Metallurgical Analysis

Report

Dominion

Engineering Report

Volume and Weight of Material Lost at Nozzle 3

51-125825-00

CRDM Nozzle Heat Treatment

25

Material Test

Report

DBNPS Reactor Vessel Head Certified Material Test Report

Intra-Company

Memorandum

Control Rod Drive Nozzle Cracking, dated May 8, 1996

Root Cause Plan Dated March 18, 2002.

Intra-Company

Memorandum

Probable Cause Summary Report for CR2002-0891, dated March 22, 2002

Meeting Minutes

DBPRC Meeting Minutes for MOD 94-0025.

Standing Order

87-015

RCS Leakage Management and Attached Policy Reactor Coolant System

Leakage Management

0620-00143210

Lukens Steel Company, Test Certificate, Chemical Analysis and Physical

Properties

Photographic Records

Picture

Vessel Head on Stand

Picture

Head and Service Structure Looking NE

Picture

Scaffolding Around Service Structure

Picture

View From Newly Cut Service Structure Manway Opening Looking into Drives

Picture

Looking Through Manway Cut in Service Structure

Picture

View of CRD Flanges Above Insulation Showing Some Removed and Some

Installed

Picture

Control Rod Drive Flanges Above Insulation

Pictures

Shielded Work Platform on Top of Service Structure

Picture

Pictures of Nozzle 2 and 3

Pictures

Area Surrounding Nozzle 3 Penetration

Picture

Nozzle 16 Quad C

Pictures

Nozzle 2

Pictures

Nozzle 3 Remnant

Pictures

From Bare Head Video Exam Conducted in 13 RFO.

Video Tape

Davis-Besse Reactor Head Inspection Under Insulation Alloy 600, 12 RFO

Video Tape

Davis-Besse 12 RFO Final Head Inspection

Video Tape

Davis-Besse Reactor Head Cleaning 11 RFO

Video Tape

Davis-Besse Weep Hole Cleaning Nozzle 67, 10 RFO

Video Tape

Davis-Besse Weep Hole Video Inspection 10 RFO

Video Tape

13 RFO Reactor Head Nozzle Remote Visual Inspection

Video Tape

Root Cause Video of Nozzle #3 and Adjacent Nozzles, March 13, 2002 to

March 14, 2002

Video Tape

PT of Nozzle #46 J-groove Weld, March 24, 2002

Potential Conditions Adverse to Quality Reports (PCAQR)

1988-0494

Condition Not Satisfactorily Resolved per PCAQ

1990-0221

CRDM Flanges #F02 and F-4 Erosion and Irregularities.

1991-0353

Boron on Reactor Vessel Head

1992-0072

CAC Cleaning

1993-0098

Boric Acid Corrosion on OTSGA Head Vent Flange

1993-0132

Reactor Coolant Leakage from CRD Flange

26

1994-0912

Documents Results of CRDM leakage Video Inspection

1994-0974

CRDM Flange Indication

1994-0975

CRDM Flange Indication

1994-1338

10 CFR Part 21 RX Adaptor Tubes

1996-0551

Boric Acid on RX Vessel Head

1996-0650

VT-2 Inspection Revealed Evidence of Leakage and Boric Acid Residue

1996-1018

IN 96-032 RV Augmented ISI

1998-0020

Inadequate Testing

1998-0649

Reactor Vessel Head Boron Deposits

1998-0767

Reactor Vessel Head Inspection Results

1998-0824

CAC Boric Acid Accumulation

1998-1164

Water in RE4597 Sample Lines

1998-1885

Found Two Carbon Steel Nuts on RC2

1998-1895

CTMT Normal Sump Leakage in Excess of 1 gpm

1998-1965

Water and Boron Accumulation on Filter Cartridges

1998-1980

Potential CAC Fouling

1998-2071

Accumulation of Boric Acid on CTMT Service Water Piping

Procedures

NG-EN-00324

Boric Acid Corrosion Control, Revisions 1, 2, and 3

PP-1102.10

Surveillance Test Procedure: Plant Shutdown and Cooldown, Revision 16

DB-OP-06903

Operations Procedure: Plant Shutdown and Cooldown

DB-PF-00204

ASME Section XI Pressure Testing, Revision 3

DB-OP-01200

Reactor Coolant System Leakage Management, Revision 3

ATTACHMENT A TO NRC AUGMENTED INSPECTION REPORT NO. 50-346/02-03(DRS)

USNRC Memorandum from J. E. Dyer to R. N. Gardner, dated March 12, 2002: Augmented

Inspection Team Charter - Davis-Besse Reactor Vessel Head Material Loss

Documented in ADAMS (Accession Number ML020730194)

ATTACHMENT B TO NRC AUGMENTED INSPECTION REPORT NO. 50-346/02-03(DRS)

NRC Briefing Slides for the Public AIT Exit Meeting Conducted on April 5, 2002

Documented in ADAMS (Accession Number ML021070811).

ATTACHMENT C TO NRC AUGMENTED INSPECTION REPORT NO. 50-346/02-03(DRS)

FirstEnergy Intra-Company Memorandum from S. A. Loehlein to H. W. Bergendahl, dated

March 22, 2002

Documented in ADAMS (Accession Number ML020860035)

March 12, 2002

MEMORANDUM TO: Ronald N. Gardner, Chief

Electrical Engineering Branch

Division of Reactor Safety

FROM:

J. E. Dyer /RA/

Regional Administrator

SUBJECT:

AUGMENTED INSPECTION TEAM CHARTER -

DAVIS BESSE REACTOR VESSEL HEAD MATERIAL LOSS

In response to preliminary information provided by the licensee on March 10, 2002, regarding

the significant loss of pressure boundary material from the reactor vessel head, an augmented

inspection team (AIT) is being sent to the Davis-Besse Plant. You are hereby designated as

the AIT leader.

A.

Basis

On March 6, 2002, during repair activities to control rod drive mechanism (CRDM)

nozzles, the licensee identified an area of wastage in the reactor pressure vessel head

surrounding the No. 3 CRDM nozzle. The licensee initially identified five CRDM nozzles

that required repairs due to cracking in the J-groove welds found during the nozzle

examinations required by Bulletin 2001-01. Wastage area in the head was discovered

when the licensee removed the No. 3 CRDM nozzle, after the penetration tube

unexpectedly moved during repair activities.

Because this was a significant unplanned degraded condition having potential generic

safety implications, an AIT was initiated in accordance with NRC Management Directive

8.3, "NRC Incident Investigation Program." The purpose of the AIT is to better

understand the facts and circumstances related to the degradation of the reactor vessel

head pressure boundary material. It is also to identify any precursor indications of this

condition so that appropriate followup actions can be taken. All followup actions

associated with the extent of condition, repairs/replacements, or corrective actions

related to plant restart will be covered through other inspection activities.

CONTACT:

John A. Grobe, Director, DRS

(630) 829-9700

R. Gardner

-2-

B.

Scope

Specifically, the augmented inspection team is expected to collect, analyze, and

document factual information and evidence sufficiently to address the following:

1.

The plant history of reactor coolant system operational leakage indications,

including trends in unidentified leakage, containment air cooler fouling,

containment radiation monitor readings, etc.

2.

The plant history of reactor vessel head material condition issues, including

control rod drive flange leakage or other sources of corrosive substances.

3.

The plant history of reactor vessel head inspection, including visual inspections,

ultrasonic testing, prior video-records of head examinations, reactor vessel head

cleaning activities, and licensee action in response to generic correspondence

for leakage and degradation of the reactor coolant system.

4.

Characterization of all reactor vessel head wastage areas, including the best

available geometric details of cavity volumes, surface conditions, surface

contaminants, etc.

5.

The probable cause(s) for the vessel head wastage.

C.

Guidance

This memorandum designates you as the AIT leader. Your duties will be as described

in Inspection Procedure 93800, "Augmented Inspection Team." The team composition

has been discussed with you directly. During performance of the augmented inspection,

designated team members are separated from their normal duties and report directly to

you. The team is to emphasize fact-finding in its review of the circumstances

surrounding the event, and it is not the responsibility of the team to examine the

regulatory process, to determine whether NRC requirements were violated, to address

licensee actions related to plant restart, or to address the applicability of generic safety

concerns to other facilities. Safety concerns identified that are not directly related to the

event should be reported to the Region III office for appropriate action.

The team will report to the site, conduct an entrance meeting, and begin inspection on

Tuesday, March 12, 2002. Tentatively, the inspection should be completed by close of

business March 22, 2002, with a report documenting the results of the inspection,

including findings and conclusions, issued within 30 days of the public exit meeting.

While the team is on site, you will provide daily status briefings to Region III

management.

This Charter may be modified should the team develop significant new information that

warrants review.

DOCUMENT NAME: G:DRS\\ML021260141.wpd

To receive a copy of this document, indicate in the box: "C" = Copy without attachment/enclosure "E" = Copy with attachment/enclosure "N" = No copy

OFFICE

RIII

RIII

RIII

RIII

NAME

JGavula:sd

CLipa

JGrobe

JDyer

DATE

3/12/02

3/12/02

3/12/02

3/12/02

OFFICIAL RECORD COPY

Retrieved from "https://kanterella.com/w/index.php?title=ML021260141&oldid=5439279"