ML20040C919
| ML20040C919 | |
| Person / Time | |
|---|---|
| Site: | Oyster Creek |
| Issue date: | 01/11/1982 |
| From: | Baunack W, Briggs L, Greenman E, Rekito W, Reynolds S, John Thomas NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I) |
| To: | |
| Shared Package | |
| ML20040C914 | List: |
| References | |
| 50-219-81-17, NUDOCS 8201290374 | |
| Download: ML20040C919 (17) | |
See also: IR 05000219/1981017
Text
__
f
r
-
..
U.S. NUCLEAR REGULATORY COMMISSION
OFFICE OF INSPECTION AND EliFORCEMENT
l
Region I
Report No.
50-219/81-17
Docket No.
50-219
C
License No. DPR-16
Priority
Category
j
--
Licensee:
Jersey Central Pcwer and Light Company
Madison Avenue at Punch Bowl Road
Morristown, New Jersey
>
Oyster Creek Nuclear Generating Station
Facility Name:
Inspection at:
Forked River, New Jersey
Intrection conducte .
August 27 - October 19, 1981
gh6(
_
Inspectors:
s
' date signed
J. A. T oma,s, Resident Reactor Inspector
l
WA
!
date signed
g S. Reynolds, Reactor Inspector
Ih
2
i
/dat'e signed
W. Rekito, Reactor Inspector
f$ $n$M
/.Q/J'7/8/
L. Briggs, R6a'ctor Inspector
date signed
Jb)ff b
-s
W. Baunack, Reactor Inspect'of
' dat'e signed
/ / f7
- Approved by:
y
<--
E. G. . Gree. man, Cnief, Reactor Projects
dats signed
Section No.'2A
Inspection Sumary: Inspection on August 27 - October 19,1981 (Report No. 50-219/81-17)
Areas Inspected:
Inspection by one resident inspector and four region based inspectors
+
plus a technical presentation by licensee management for senior NRC regional manage-
ment to review the events relating to tube failures in two shutdown cooling heat
exchangers. The inspection included a review of the sequence of events, safety
verification, evaluation of failure mechanism, evaluation of repairs, review of
corrective action, and review of applicable documents and procedures.
Results: Violations: one (Failure to make a report pursuant to 10 CFR 50.72 - detail 3).
8201290374 820115
PDR ADOCK 05000219:
G
Region I Form 12
(Rev. Ay il 77)
'
__
-
-
,
,
..
,.
,
e
OETAILS
1.
Persons Contacted
R. Arnold, Senior Vice-President, JCP&L
K. Bass, Engineer
J. Carroll, Director, Oyster Creek Operations
P. Clark, Vice-President, JCP&L
C. Cowfer, Manager, Materials Technology
D. Croneberger, Director, Engineering and Design
J. -De Blasio, Engineer
K. Fickeissen, Manager, Plant Engineering
I. Finfrock, Chairman, General Office Review Board
F. Giacobbe, Materials and Welding Manager
.
E. Growney, Safety Review Manager
R. Keaten, Directc , Systems Engineering
J. Kneuble, Manager, BWR Licensing
M. Laggart, Licensing Supervisor, Oyster Creek
R. Lorenzo, Superviscr, Engineering Projects
T. Quintenz, Engineer
A. Rone, Engineering Manager
J. Sullivan, Manager, Operations
J. Thorpe, Director, Licensing and Regulatory Affairs
P. Walsh, Manager, Plant Analysis
i
,
The inspectors also interviewed other licensee personnel during the
course of the inspection including management, clerical, maintenance
and operations personr.el .
2.
Introduction
On August 27 and 28,1981, tube failures occurred in two of three
shutdown cooling system (SDCS) heat exchangers while in service for
reactor decay heat remeval. The SDCS is a closed loop system designed
to remove decay heat from the reactor during shutdown operations by
circulating water from the 'E' recirculation loop through the heat
exchangers. The system consists of three pump and heat exchanger
combinations which share common supply and return lines connected to
the 'E' reactor recirculation loop.
Reactor water. flows through the
heat exchanger tubes and is cooled by circulation of Reactor Building
-Closed Cooling Water (RBCCW) on the heat schanger shell side. The
system provides only decay heat removal for normal shutdown operation
and performs no accident mitigating functions.
Failure of two heat
exchangers in such a short time period caused doubt about the integrity
of the remaining heat exchanger.
In addition, orie possible cause of
the failures was thought to be corrosion due to salt water contamination
!
- -
.
t-
. - . .
-
,,
._
m
- .
.
s
_..
.
.
3
.
-
of the RBCCW system. This caused concern for the integrity of
other stainless steel components cooled by RBCCW. Additional
component failures could lead to loss of ability to keep the -
reactor in a cold shutdown condition, and inability to maintain
fuel pool cooling. TFe sequence of events leading to the SDCS
heat exchanger failures, safety concerns created by the failures, and
-licensee corrective actions are discussed in detail in this report.
<
3.
Sequence of Events
On Augult 27,1981, the reactor was in a cold shutdown condition
~ with the 'C' SDCS heat exchanger in service maintaining reactor
0
temperature between 1600 and 180 F.
At about 11:35 a.m., a decrease
in reactor water level was noted when one of four low water level
sensors actuated, causing a low water level half scram to be annunciated
through the reactor protection system. The control room operator
observed a simultaneous step increase in RBCCW activity and
annunciation of the RBCCW surge tank high level alarm. The operator
immediately isolated the SDCS, stopping the decrease in water level.
Reactor water level was restored to normal and the ' A' SDCS heat
exchanger was placed in service. During this event, reactor water
level dropped from 5.7 feet to 3.8 feet as indicated on the GEMAC
narrow range recorder in the control room. This drop in water level
occurred in about 10 minutes and is equivalent to a reactor leakage
rate of 400 gpm. The RBCCW system activity increased from about
1E-4 to 1E-3 microcuries per milliliter gross gamma, but no increase
in background radiation or airborne activity was detected. No release
of radioactivity to the environment occurred.
The 'A' SDC loop was used to cooldown the reactor to about 1350F.
The SDC system was then secured to perform a test to determine the
feasibility of cooling the reactor using the Reactor Water Cleanup
System (RWC'JS) nonregenerative heat exchanger (NRHX). At 8:45 a.m.,
on August 28, 1981, the 'A' SDCS heat exchanger was returned to
service with the reactor temperature at 1900F. At about 3:35 p.m.,
with the reactor temperature being maintained between 1600 and
1800F, the RBCCW surge tank high/ low level alarm annunciated. Addi -
tionally, an individual on the 95 elevation of the reactor building
reported that the surge tank was overflowing, and a slight' drop in
i
reactor level was noted. These parameters were indicative of a tube
,
leak in the 'A' SDC heat exchanger. The SDC-system was isolated and
'
the leakage stopped. The licensee decided not to place the 'B' SDC heat
exchanger in service and began planning for alternate methods o' decay
heat removal,
t
t
I
e
. . -
.
m...
. _ - - - . _ , . . . _ , _ - . . , ,
.
. _ - - - - - - , _ _ . ,
_ - , - _ _ _ _ - . _ - - - . -
.
.
-
,. -
- - . .
-
.
.
-
_
.
-
.
.
-.
1
.
4
The licensee mobilized the management and technical staff in an alert
!
status and notifieA the required government agencies in accordance with
,
the Emergency Plan. At about 9:30 p.m. on August 28, procedure 203.3,
'
" Alternate Shutdown Cooling Method", had been reviewed / approved by the
i
Plant Operations Review Consnittee (PORC) and implemented. The procedure
provided for reactor cooldown by operating the Reactor Water Cleanup
Systam (RWCUS) at a flow rate of about 500 gpm and a letdown rate to the
condenser hotwell of about 300 gpm. A condensate pump was placed in
service to maintain reactor water level between 155 and 180 inches above
the top of the active fuel. The system was placed in service at a reactor
temperature of 152 degrees F and was able to maintain a cooldown rate of
,
!
about 1.5 degrees per hour.
A review of events preceeding the failure of the SDC heat exchangers found
that there had been a previous _ tube failure in an RBCCW heat exchanger
that had. allowed salt water contamination of the RBCCW system from the
_
.
service water system. On August 11,_1981, a four gpm leak from the
.
RBCCW sytem was detected when a recently installed flow integrator on
'
the make-up line was placed in service. Heat load on the system prohibited
isolation of the heat exchangers to detennine the location of the leak.
On August 19, 1981, the leak was found to be from the west RBCCW heat
exchanger into the Service Water System (operating with a pun'p discharge
pressure of about 60 psig.) The heat exchanger was isolated and the RBCCW
system was operated in a one heat exchanger /two pump mode.
.
On August 23, the licensee decided to secure one RBCCW pump due to fu ther
f
i
t
reduction in the system's heat load. The heat exchanger bypass valve was
throttled closed to increase system pressure to about 105 psi and to direct
4
more flow through the heat exchanger in preparation for securing one pump.
j
When flow was increased, tube vibration was noted .in the heat exchanger.
-
One pump was secured and system pressure dropped to about 70 psi which is
lower than nonnal operating pressure. The pump was restarted to restore
pressure. Following restart of the idle pump, the RBCCW surge tank
level dropped rapidly indicating a failure of a heat exchanger tube.
The east heat exchanger was taken out of service and isolated, and the
west heat exchanger was placed back in service. A water sample from the
RBCCW system was analyzed and found to contain 72 ppm Nacl (about 45 ppm C1).
Salt water had apparently leaked into the system during the transient. An
'
ion exchange column was put in service on the system to remove the salt
i
'
contamination. On August 26, the failed tubes in the east exchanger were
repaired and the unit was returned to service.
'
l
During the period that the system was operated with a 4 gpm leak and
I
during the pump. realignment evolution when gmss RBCCW tube failure
occurred, a release of RBCCW water to the environment via the-Service
Water System occurred.
l
I
r
I
i
-
. .
.
.
.
-~~
,,
- , _ _
~ .
r
-
..
5
The RBCCW system operated from August 11 to August 19, 1981, and again
from August 23 to August 26, 1981 with a known 4 gpm leak through the
west heat exchanger. This leakage was to the service water system, which
operates at a pressure about 40 psi lower than the RBCCW system pressure.
The service water systerr discharges to the cooling water. discharge canal,
thus, the leakage from the RBCCW system was ultimately to the environment.
The RBCCW system activity levels average about 1.1 E-4 microcuries per
milliliter, therefore, a total release of about 27,000 microcuries total
activity was discharged to the environment darIng this time period. The
licensee did not recognize the RBCCW 1eakage as an unplannM release of
radioactivity and no report was made to the NRC pursuant to i0 CFR CO.72.
Failure to report such a release is in noncompliance with 10 CFR 50.72
(219/81-17-01).
4.
Operational Safety Verification
4
a.
rollowing failure of
'C' then 'A' SDCS heat exchangers, the
.
licensee determined that it would not be advisable to place the
'B' heat exchanger in service until the cause of the failures
i
was known. As a result of this de-ision, the normal methods of
decay heat removal while in cold shutdown were not available. At
i
i
about 4:00 p.m. on August 28, the inspector toured the control room
and verified that adeauate safety systems were available for
>
i
decgy heat removal . The reactor was at a temperature of about
140 F with a gradual heatup rate. The control room operators
had secured all but the 'B' reactor retirculation' pump to
j
minimize the heatup rate and had left the suction and discharge
valvas of the 'B'
and 'D' loops open to insure adequate flow
between the annulus and core rec'ons of the reactor vessel. The
licensee decided to. leave one rec.rculation pump in operation to
prevent stratification ir, the core region and subsequent loss of
core temperature monitoring capability. The Core Spray System
and Isolation Condenser System were available if needed; and the
steam and condensate systems were available for decay heat removal
by drawing steam to the main condensers. The licensee technical
staff proposed the following alternate methods of decay heat
removal:
(1)
Circulate reactor. water at the maximum rate through
the reactor water cleanup system (RWCUS) and use the
nonregenerative heat exchanger for cooling. This
l
method was feasible and normal RWCUS operating
procedures were adequate but only marginal cooldown
rates could be achieved.
'
(2)
Circulate reactor water at the maximum rate through
the RWCUS and discharge water from the system to the-
condenser hotwell through the normal letdown path.
This method would allow additional heat removal by the
main condenser and would require operation of a
condensate pump to feed water back to the reactor.
1
w
s.-
e
,
.
v-r
,
,
-w
,
y
- ,e--e
w.-1y__
c
.
..
'
.
6
(3)
Flood the reactor vessel to the main steam lines and
allow reactor water to drain throuah the steam lines
to the main condenser. The water would be cooled by
normal operation of the circulating water system and
fed back to the reactor through the condensate system.
This method would require preparation cf special
procedures and a detailed review of the structural
design of the main steam system to det. ermine if the
piping supports could carry the weight of fully flooded
steam lines.
(4)
Flood the reactor vessel and the isolation condensers,
operate a recirculation pump and force circulate
reactor uter through the iso-condensers. This
method would require considerable procedural review,
development of a feed and bleed system for the shell
sides of the iso-condensers, and analysis to determine
the heat removal c::pability in this mode. This
method was later determinednot to be feasible.
(5)
Allow the reactor to repressurize and remove decay
heat by steaming to the main condensers or isolation
condensers.
Existing procedures were adequate to use
either of these methods for decay heat removal but
would require coming out of the cold shutdown condition.
The licensee chose to use method (2) above as the primary means
of decay heat removal. Procedure 203.3, " Alternate Shutdown Cooling
Method" was implemented at about 9:30 p.m. on August 28,1981.
The conditions in the RBCCW system at.the time of the SDC system
failures were such. that chloride stress corrosion cracking of
stainless steel components was possible. This created concern for
the integrity of other components in the system. Of immediate
concern were the reactor recirculation pump coclers, the RWCUS
nonregenerative heat exchanger (NRHX), and the fuel pool coolers.
The licensee needed to develop contingency plans for core decay heat
removal and fuel pool cooling in the event any of these critical
components failed.
'
The alternate shutdown cooling procedure was written to include-
contingencies for failure of the NRHX and loss of the reactor
recirculation pumps.
If the NRHX tubes failed, cooling would
continue through the same flow path but the RBCCW side of the NRH.X
would be-isolated and cooling would be accomplished in the main
condenser only.
If the recirculation pump coolers failed, the
l
.
.
,.
7
reactor temperature would be allowed to increase to the point
of steaming to the main condenser. The licasee prepared procedures
for the operation of the augmented fuel pooi cooling system to
be.used in the event of fuel pool cooler failure. The augmented
fuel pool cooling system was installed in 1975 to provide sufficient
fuel pool cooling to accommodate a full core off-load into the
fuel pool. The system consists of two plate type heat exchangers
and a circulating pump on the 75 level of the reactor building and
can be valved into the existing fuel pool system piping.
It
was necessary to install a temporar; p wer feed from the 'C'
pump breaker to the augmented fuel pw. cooling pump and to
' jumper out the SDC pump valve interlocks. This system was made
operational and a contingency procedure was developed to provide
a feed and bleed method of cooling in the event of a complete
loss of the RBCCW system.
The inspectors reviewed the licensee's procedures and attended
technical planning meetings and were satisfied that adequate
methods and contingency plans had been established to insure
adequate core cooling without the use of emergency core cooling
systems.
It was later determined as discussed further in this report that
the heat exchanger failures were unrelated to chloride stress
corrosion.
b.
On August 29, 1981, the licensee attempted unsuccessfully .to
isolate and drain the SDC heat exchangers. Leakage past the
RBCCW valves made draining of the shells impossible. The licensee
0
planned to cooldown the reactor to 125 F, establish alternate
fuel pool cooling, and then shutdown the RBCCW system so the SDC
heat exchangers could be isolated by the installation of blank
flanges. The inspector reviewed the procedures for the evolution
and found that the intent was to cooldown the reactor and secure
the recirculation pumps during the RBCCW stoppage.
In_the eveni.
of reactor heatup, the licensee intended to draw a slight vacuum
on the reactor through the main steam lines with the mechanical
vacuum pump and steam to the main condenser to remove decay
heat. The inspector expressed concern for natural circulation
cooling of the core at sub-atmospheric pressure. The core is
analyzed for natural circulation cooling during power operation and
<
for cooling at pressures down to atmospheric pressure. No
analysis is documented for sub-atmospheric conditions. The
inspector stated that a safety e'vuluation pursuant to 10 CFR 50.59
must be performed to ensure adequate core circulation to provide
adequate cooling and core temperature monitoring. The licensee
performed a safety review and submitted the results t; NRC:RI for
review. On September 2,1981, the safety review was submitted and
the RBCCW stoppage and SDC isolation procedures were completed without
incident.
.
-,
.-
-,
. - , -
,
,
.
-
_
.
..
8
5.
Evaluation of Tube Failures
a.
Description of Affected Units:
The three SDC HX units in question were manufactured by South-
western Engineering Company (SWECO), and are SA 213 type 304
(seamless) "U" tubed horizontal units with 367, 3/4" x 18 BWG
AVERAGE (0.049 + 10%) wall tubes on a 15/16" pitch. The tubes are
welded to a. stainless clad tubesheet. The units are designed as
Section III Class C on the tube side and Section VIII on the shell
side, and are designed for 150 psig/350F on the shell side and
1250 psig/350F on the tube side. The materials are carbon steel
except the tubes and tube sheet cladding which are austenitic
stainless steel. The tubesheet is approximately four inches thick.
The tube support plates are partial " half baffles" with each tube
leg supported by two tube support plates (TSP). There are no
antivibration bars at the "U" bend area. SWEC0 drawing DM 77680-
shows the straight length of the tubes to the point of tangency
to be 76" for the short radius tubes and 79 3/4" for the largest
radius tube. The flow on the shell side of the unit enters at
the back cf the shell on the bottom near the "U" bends. SWEC0
drawing DM 77680 shows the last tube support plate to point of
tangency of the bend dimension (for the isrgest radius bend) to be
1 3/16" for the bottom legs of the tubes. The tubes are welded to
the clad tubesheet and post weld rolled (3-5% wall reduction) for
two inches starting one half inch behind the tube weld. Valve
leakage on both the primary and RBCCW sides of the units has
caused the units to become pressurized during normal reactor operation.
b.
Failure Location:
,
The licensee's intention was to conduct a shell-flooding static
hydro to locate failed tubes in "A" and "C"
units; conduct
Eddy Current (EC) inspection of a portion of the tubes (47 minimum);
and remove failed tubes for metallurgical examination. The EC
procedure used was NES CONAM 42-EC-021 which was a SC XI
procedure with calibration to Article IV-3000. The calibration stan-
dard tube contained the regular drilled hole artificial defects
plus 4 OD grooves and 1 1D groove. The calibration sample was SA 249
type 304 and the single test frequency was 300 MHZ.
The licensee using CONAM personnel inspected a pattern of tubes on
units "A" and "C".
The pattern selected looked at peripheral
tubes and small radius bend tubes. The inspector witnessed the EC
testing of the tubes in HX "C" and a re-testing at higher
sensitivity level of HX
"A".
Borescopic examination indicated
probable circumferential tube failures at or near the second tube
.
-
_
~
.
9
support plate in units "A" and "C"
on the same tube, i.e.
R31 C68.
Visual examination of unit "A" indicated a plugged
tube adjacent to the recently failed tube, i.e. R31 C64. This
tube had been plugged with an Elliot driven plug.
When the Elliot plug was removed, the tube weld appeared to be
cracked. Adjacent tube welds were given a PT test with no
evidence of additional T/TS weld cracking. The " cracked weld"
was later determined to be sound. The licensee attempted to " pull"
the failed tube in unit "C" (R31 C68) but inadvertently pulled an
adjacent tube (R31 C64) which had not failed. Approximately
12 inches of the tube was pulled through the tube sheet before
realizirig that the wrong tube had been pulled. Tube R31 C68 was
later pulled but was thought to have broken off while pulling.
Later examination determined that a measurement error had been
made upon pulling and the entire section of the failed tube had
in fact been removed. The inspector examined the fracture face
of the tube and observed the following:
1.
The tube support plate ring from rotation of the tube
support plate hole was observed. The location of the
circumferential failure was adjacent to the tube
support plate (TSP) on the "U" bend side.
2.
The failure can be categorized as a " brittle failure"
of a ductile material. There was essentially no
evidence of plastic deformation associated with the
fracture.
3.
The fracture face showed no immediately apparent " fresh"
fracture face. The entire fracture face was covered
with sufficient corrosion products to appear like an
old fracture.
The fracture face from tube R31 C68 was sent to a metallurgical
laboratory and the results of the analysis are documented in
GPU NUlear Materials Technology Laboratory Report 80128. This
report indicates the rost probable cause of failure to be fatigue.
The proximity of the. failure location to the tube support plate
left the single frequency EC test in question for this general
location. (A multi-frequency technique was later required for
detecting defects within the magnetic disturbance area of the
tube support plate).
c.
Evaluation of Stress Corrosion Cracking Potential:
The inspector analyzed the corrosion characteristics of the SDC
HX units. The heat exchangers must be analyzed in accordance
with their actual service conditions rather than the design
conditions.
,
.
.
_
'
~
.
10
The valves on the tube side and the shell side leak. When the
plant is operating, the valve leakage produces a 1000 psig
pressure on the tube side with the shell full of water which will
build up to the pump pressure of approximately 100.psig. The
heat flux direction in this mode of operation has the highest
temperature on the tube side. The water in the shell will be
essentially stagnant with only a convection current due to heating
and possibly steaming of the water on the shell side. The tube
to tube hole crevice will be a concentrating crevice due to the
higher heat on the tube side. This condition (RV power operation)
produces the highest service induced hoop stress.
~
Assuming the thinnest wall and the largest ID, the service induced
hoop stress is 6735 psig. The minimum yield strength f5r SA 213
Grade 304 tubes is 30,000 psi, but the expected yield strength is
35-50,000 psi (nominally 35-40,000). The fabrication (straightening)
residual stresses could be greater (possible YS level) than the
service induced stresses. Whea the unit is being operated as a
cooler, the pressure difference across the tube wall was estimated
to be 65 psig. This would result in an applied hoop stress of
less than 500 psig.
The only two methods of producing a " brittle" failure in a
ductile material such as the 304 stainless steel tube are by
fatigue or SCC. The most probable SCC mechanisms affecting the
304 are IGSCC and chloride ion SCC.
IGSCC would normally be
expected to be an ID to 0D failure mechanism on the oxygenated
side of the unit. This mechgnism ig temperature dependent with
The mechanism also
maximum failure rates at 200 C.(392 F).
requires a tensile stress and a sensitized pertion of the tube.
The only possible sensitized areas are the area adjacent to the
T/TS weld and the HAZ from the electric resistance return bend
solution annealing operation. The stress pattern in the T/TS
weld area could give longitudinal and/or circumferential cracks,
but the crackt would be short and limited to this area. The
stress pattern from the solution annealing of the bends could
give a similar limited crack pattern. The EC testing inspected
the area of the return bend which would contain any sensitized
microstructure from solution annealing but could not inspect
behind the T/TS weld. There is no evidence to suggest failures
within the tubesheet.
IGSCC can be ruled out as a failure mode.
Assuming that chloride ion SCC were a problem, the chlorides would
affect the OD of the tubing. The reported chloride concentration
due to intrusion of salt water was 75 ppm (as NaCli. The tube
metal temperature during exposure to C1 was estimated to be 140-
160F. SCC is temperature dependent with minimal effects even
with high chloride levels below 150F. The tensile stress on the
OD of the tubes would be caused by the mechanical expansion (rolling)
transition from the rolled tube to the unrolled tube (this is within
.
.
=
w
.
--
.
.
..
11
the tubesheet) or by unstress-relieved return bends. The only
other significant tensile stress in the tubing would be the
possible residual stress from straightening operations. The
roll transition stress pattern is complex, but the tensile stresses
run out so there could only be short cracks (if they did occur)
with minor leakage. The hoop stress from tube straightening
would result in a longitudinal crack in the tubing.
It would stop
at the solution annealed bends at the HAZ from the thermal process.
This type of cracking should be readily apparent by single or
multiple EC testing. Unstress-relieved return bends would
i
preferentially fail at the small radius bends.
Evaluation of
the data indicated the tube failures were not caused by chloride
^
.
d.
Evaluation of SDC HX Tubing Reliability:
The licensee's. program to evaluate the reliability of the SDC
HX's tubing consisted of static shell side flooding hydrotest,
single frequency and multi-frequency ID probe eddy current
testing, borescopic ID examination of selected tubes, macro and
micro metallographic analysis of selected tube samples, and shell
side and tube side hydrostatic tests.
The intent of the eddy current testing was to determine the axial
location of through wall failures and to inspect the tubing for
partial wall failures.
Due to lack of imediate availability of
multi-frequency eddy current (EC) equipment, the licensee
initially conducted tests with single frequency EC equipment. A -
sampling pattern of tubes was selected which maximized inspection
of Row 31 (and other peripheral tubes), the small radius bend
tubes and tested a few tubes in the body of the tube bundle.
Approximately 10% of the tubes were single frequency EC tested in
Units A and C.
All tubes were tested with a 300 KHZ differential-
coil technique and approximately 5% of the tubes were also tested
with an absolute coil technique. The tests showed no reportable
indications with the differential coil and no reportable wall
thinning. The failure location was determined (physically) to be
adjacent to tube support plate #3 (bottom leg closest to the U bend
end ) . The proximity of the failures to the tube support plate
indicated that if this were the only failure location, single
frequency techniques could not be used for partial tube wall
defect detection (due to tube support plate perrneability).
Once the axial location on tube R31 C68 in the C Unit was determined,
the licensee attempted to remove this tube for metallurgical
evaluation. Due to a communications error, the licensee started to
remove tube R31 C64 rather than R31 C68. When approximately 12 inches
of R31 C64 was pulled through the tubesheet, the licensee discovered
the error. Attempts to remove the defective tube (R31 C68) in Unit C
were unsuccessful apparently due to tube / tube hole interference within
the tube sheet.
.
r
-
,
,
.
.
12
The licensee obtained multi-frequency EC testing equipment and
tested sample tubes in all three SDC HX's. The equipment employed
2 - 300 KHZ and 1 - 75 KHI differential coils and one 10 KHZ
absolute ' coil . Analysis was performed and reviewed by qualified
EC testing experts. The tubes (entire length) tested were as
follows:
R0W
TUBES
A
B
C
31
72, 68, 64, 60, 58,
X
X
X
54, 50, 46
29
82, 78, 74, 70, 66
X
X
X
.
62, 59, 44, 40, 36
'
29
56, 52, 48
X
X
27
88, 84, 34, 30
X
X
X
27
80, 76, 72, 68, 64
X
X
60, 58, 54, 50, 46
42, 38
25
90, 28
X
X
X
23
96, 92, 26, 22
X
X
X
21
98, 20
X
X
X
19
104, 100, 60, 58, 14
'X
X
X
17_
106, 59, 12
X
X
X
15
108, 60, 58, 10
X
X
X
13_
110, 90, 86, 82,
X
X-
X
78, 74, 44, 40,
36, 32, 28, 8
>
11
112, 6
X
X
X
9
11 0.8
X
X
X
7-
112, 6
X
X
X
5
114, 4
X
X-
X
.
,,
-.
4
-
-
_ . ~ , .
- - - . ,
,
,
. - - , .
.~e.
,-
,
.
..
.
13
3
116, 60, 58, 2
X
X
A
1
114, 94, 74, 44,
X
X
X
24, 4
Note:
R31 C64 "A" unit terted af ter removal of the plug.
The analysis of the multi-frequcrey testing indicated non-rejectable
" dings" and " unresolved luw '(vel indications" in all three units.
Only one significant indication was noted in Unit A and none were
noted in Unit C.
The siunificant indication in Unit A was
reported as a "cracklike'(circumferential) indication" estimated
to be equivalent to a 63% through wall crack in tube R31 C46 located
at tube support plate #3. Analysis of the data obtained from
Unit B resulted in 17 reportable indications. These indications
were reported as " partial through wall cracklike indications".
The indications were all, but two, in the bottom leg of Row 31.
The indications are reported at tube support plates, predominantly
tube support plate #3. The circumferential "cracklike
indications" are reported to vary from 31-60% through wall. Two
tubes in Row 29 (R29 C70 and R29 C78) are reported to have crack-
like indications at tube support plate #3.
The borescopic examination of the failed tubes (R31 C68 in Units
A & C and R31 C64, Unit C) indicated the fracture faces to be
located at TSP #3 on the bottom legs. Borescopic examination of
other tubes in areas not known to be failure areas resulted in
data of questionable value due to anomalus appearance of ID
films.
e.
Evaluation of Additional Tube Sections:
The following 3 " long tube segments were removed for visual
inspection:
-- Tube R31 C46 at TSP #3.
-- fube R31 C46 at TSP #4.
-- Tube R31 C64 at TSP #4.
These tube segments were chemically decontaminated prior to visual
inspection. Tube R31 C46 at TSP #3 was previously reported to have
a 63% through wall EC "cracklike indication" . Visual inspection
of these tubes by the NRC inspector indicated the following:
.
.
.
14
The tubes were rotating in an elliptical pattern in
--
the TSP and had burnished areas in a pattern eccentric
to the OD cf the tube.
The burnished area appeared to be severly cold worked
--
on the tube OD surface.
-- The edge of the burnished area on at least one of the
tubes showed a worn-in or plastically deformed
circumferentially oriented ridge.
-- There was no visually observable crack in tube R31 C46
at TSP #3.
f.
Previous Failures of SDC HX Tubing:
The NRC Inspector reviewed the documents for Maintenance Job
Order 7158 QASL 964 dated 3/1/77 written to locate and plug
leaking tubes in SDC Unit A.
Procedure 713.1.002, Revision 1,
dated 2/19/76 was used for inspection and repair. Tube
R31 C64 was found leaking on the bottom leg and was plugged with
an Elliot plug. The axial location of the plug was not determined.
The Elliot plug was removed from tube R31 C64 and the location of
the failure was determined to be at TSP number 3.
The SDC HX's at Nine Mile Point, Unit 1 are duplicates of the
Oyster Creek units, and were also manufactured by SWECO. Two
of the three units each had one tube found to be leaking on tests
conducted on an outage in March and April 1977. Coincidentally,
the tubes that leaked were the lower leg of Row 31. The axial
leak location was not determined. These tubes were plugged with
tapered driven plugs. The units have been operable since the tube
plugging.
6.
Post Repair Test Activities
a.
Scope and Acceptance Criteria:
'
The inspector reviewed the Special Test procedures and test
results data to ascertain that the leak-tight integrity of the
repaired Shutdown Cooling System Heat Exchangers was adequately
demonstrated. The acceptance criteria for this review included
the facility Technical Specifications, ASME Boiler and Pressure
Vessel Code, and inspector judgment.
b.
Findings:
A
,-
15
With the exception of the items below, the objective of demonstrating
the leak-tight integrity of the repaired heat exchangers was
accomplished and the inspector had no further questians in this area.
1)
Technical Specification 6.15, Integrity of Systems
Outside Containment, requires leakage tests be performed
for the Shutdown Cooling System.
In response to
=
NUREG 0578, item 2.1.6a, the licensee comitted to
performing periodic inter-system leakage tests
past heat exchanger tube boundaries, as part of the
ongoing Leakage Reduction Program.
The inspector noted that the Leakage Reduction Prccedure
665.5.009 did not address the possible leakage past
.
heat exchanger tubes but was satisfied that the commitment
was met at this time by the conduct of Special
Procedure 81-135. The 1icensee representative
acknowledged that the existig formal leakage reduction
program procedures do not check. inter-system leakage
past heat exchanger tubes. He stated that applicable
procedures would be changed -to satisfy this comitment.
The inspector stated that this matter is unresolved
pending NRC review of revised leakage reduction program
procedures for the Shutdown Cooling System. (219/81-17-02).
2)
The Safety Evaluation of Tube-Plugging Methods stated
that installed mechanical plugs leak-tight integrity would
be demonstrated by a heat exchanger shell-side pressure
test with no visible leakage as the acceptance criterion.
The inspector noted that the purpose of both Special
Procedure 81-114 and 81-135, " Shutdown Cooling Heat
Exchanger Tubeside Hydrotest" was to demonstrate the
leak-tightness of tubes and installed mechanical plugs.
The licensee representative acknowledged this omission
and stated that the Safety Evaluation would be
revised to include demonstrating leak-tightness of the
mechanical plugs in both directions with no visible
leakage as the acceptance criterion.
7.
System Operating and Emergency Response Procedure Changes
The inspector reviewed revisions to procedures associated with the Shut-
down Cooling and Reactor Building Closed Cooling Water Systems which
resulted from the heat exchanger failure events. The acceptance
criteria for this review included Technical Specifications,
,
applicable regulatory guidance, and inspector judgment for adequacy of
.
r
.
. - .
. __
,
.
16
measures established to;
1) prevent system operation in the manner
believed to be the cause of failures; 2) detect similar failures as
soon as possible; 3) respond to similar failures by bringing the
plant to a shutdown condition in a safe and orderly fashion.
As a result of this review, no unacceptable conditions were identified
and the inspector had no further questions at this time.
8.
Document Review
The following documents were reviewed during this inspection:
-- Procedure SP 81-111, Failure of RBCCW to Recirculation Pump Seal
Coolers and Motor Bearing 011 Coolers
Procedure SP 81-112, Augmented Spent Fuel Pool Cooling System
--
-- Procedure SP 81-113, SDC HX Shell Side Isolation
-- Procedure SP 81-114 SDC HX Isolation, Tube Inspection and Repair
Procedure SP 81-116, Alternate Means of Removing Reactor Decay
--
Heat Utilizing the Main Condensers
Procedure SP 81-117, SDC HX Tube Plugging
--
Procedure SP 81-118, Fuel Pool Cooling System Feed and Bleed
--
-- Procedure SP 81-119, RBCCW System Removal and Contingencies for
Addition of Line Blanks
-- Procedure SP 81-120, Recirculation Pump Operation Without RBCCW
-- Procedure SP 81-121, NDE of Shutdown Heat Exchangers
Procedure SP 81-122, Tube Removal from Shut Down Heat Exchanger
--
A and C
Procedure SP 81-127, Opening the Shell Side of SDC HX
--
Procedure SP 81-129, Re-Installation of the Shutdown Cooling Heat
--
Exchanger NUO1-A Shell Dished Head
Procedure SP 81-134, Shutdown Cooling HX Isolation Valve Leak Test
--
-- Procedure 5P 81-135, Shutdown Cooling HX Tubside Hydrotest
-- Procedure SP 81-136, Shutdown Cooling HX Shell Side Valve Replacement
Procedure 203-3, Alternate Shutdown Cooling Method
--
Procedure 305, Revision 12, Shutdown Cooling System
--
Procedure 530, Pavision 3, Loss of the Reactor Shutdown Cooling System
--
Procedure 515.4, Revision 5, Small Pipe Break Outside Drywell
--
Procedure 309.2, Revision 4, Reactor Building Closed Cooliag Water
--
System
-- Procedure 507.1, Revision 7, Reactor Building Closed Cooling Water
System Failure
Procedure 665.5.009, Revision 0, Reactor Shutdown Cooling Leakage
--
Reduction Procedure
SWECO Dwg DM 77680, Revision 3, Bundle Assembly
--
SWECO Dwg DM 77367, Revision 6, Shell
--
-- SWECO Dwg DM 77678, Revision 3, Head and Cover Details
SWECO Dwg DM 77683, Revision 1, Gasket
--
.
~
-
.
~
-
.
.
..
17
-
SWECO Dwg M 77681, Tube Sheet Drilling Template
SWECO Dwg M 77676, Revision 2, Tube Schedule
--
SWECO.Dwg S-247, Thimble Plug Details
--
EWECO Dwg FS-452, Solid Plug Details
--
SVECO BT-8TS-M-TIG, WPS for T/TS Welding
--
-- SWICO Spec Sheet, Job 5-5026, File 26496, SDC HX type 6C-42-72H
- , Job Order 7158, QASL 964, dated March 1,1977, SDC HX Tube Leak
Inspection
GPU Nuclear Materials Technology Laboratory Report 80128, Preliminary
--
Analysis of the Oyster Creek "A" SDC HX Failure, dated September 28,
1981
GE SDC HX Equipment Specification 21A1507, Revision 0, dated
--
October 26, 1964
GPU Nuclear Interoffice Memo dated July 3,1981, RBCCW System Review
-
--
-- CONAM EC Test Procedure 42-EC-021
CONAM EC Test Procedure 42-EC-043
--
Letter from JCP&L to NRC dated January 4,1980, NUREG 0578
--
Implementa tion
-- Safety Evaluation of tube plugging methods for Oyster Creek Shutdown
Heat Exchanger Failures
Safety Evaluation of using the main condensers to remove reactor
--
decay heat to maintain reactor coolant temperature less than 2120F.
9..
Conclusions
The most probable cause of the shutdown cooling heat exchanger tube
failures is flow induced vibration resulting in circumferential
fatigue failures adjacent to the third tube support plate. Although
this cannot be verified by metallurgical evidence,-it best fits all
of the existing data. Modifications to the RBCCW system have provided
improved flow monitoring instrumentation. The plugged tubes will
act as flow impingement baffles for the other tubes and revisions to
system orerating procedures should preclude operation at flows that
would cause tube vibration.
Repairs to the 'C' SDC heat exchanger were impractical, partially due
to potential damage to the tube bundle caused by the " pulling" of an
intact U-tube. The remaining units will provide adequate decay heat
removal capability to maintain the reactor in cold shutdown under
worst case conditions.
10.
Unresolved Items
Unresolved items are matters about which more information is required in
order to ascertain whether they are acceptatle items, items of
noncompliance, or deviations.
The unresolved item identified during
this inspection is discussed in paragraph 6.
.
.
,
-
. _.
3. ,
-
..
-
. --
.
.
.
s
,
,
J
'
p'
.
18 is
,
p:
.
,
.
'll. : Technical Presentation
'
-
-GPU a'nd JCP&L Corporate and site personnel presented data, evaluations
and ccnclusions regarding the events related to failures of tubes in
two shutdo'wn cooling heat exchangers including proposed corrective
' actions. tir. P.- R. Clark, Vice President; Nuclear and members of his
staff and Mr. J. M.tAllan, Deputy Regional Administrator, Region I and
members of the Region I staff participated.
~
.. . . . . .
_.
. . -
- ., ~ . _
--
'
12.
Exit Interview
'
- At piriodic intervals during the inspection, meetings were held with
senior facility and corporate management to discuss inspection scope
and. findings. -In addition,' a meeting between senior NRC:RI and licensee
manasement was held on September 10, 1981 to discuss the sequence of events
leading to the SDC heat exchanger failures, current plant status, and
licensee plans for maintaining the plant in a safe condition.
.
.
.
-
.-
-