ML20151U618
| ML20151U618 | |
| Person / Time | |
|---|---|
| Site: | Catawba  |
| Issue date: | 04/13/1988 |
| From: | Peebles T NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II) |
| To: | |
| Shared Package | |
| ML20151U552 | List: |
| References | |
| 50-413-88-14, 50-414-88-14, NUDOCS 8805020106 | |
| Download: ML20151U618 (50) | |
See also: IR 05000413/1988014
Text
4
1 ' 2frIo '
- .6
UNITED STATES -
.['
4 'o -
k0
NUCLEAR REGULATORY COMMISSION
REGION 11 -
a4 101 MARIETTA STREET, N.W.
D f ATLANT A, G EORGI A 30323
%,...../
U.S. NUCLEAR REGULATORY COMMISSION
REGION II
AUGMENTED INSPECTION TEAM-
-REPORT NOS. 50-413/88-14 AND 50-414/88-14
Licensee: Duke Power Company
422 South Church Street
Charlotte, NC 28242
Docket Nos.: 50-413 and 50-414
License Nos.: NPF-35 and NPF-52
Facility Name: Catawba Units 1 and 2
Inspection Conducted: March 10-18, 1988
Team Members: B. R. Crowley, Reactor Engineer, RII
K. N. Jabbour, Catawba Project Manager, NRR
P. S. Lam, Chief, Reactor Systems Section, AE00
M. S. Lesser, Resident Inspector, Catawba
W. T. Orders, Senior Resident Inspector, McGuire
.,
Team Leader:
T.
/
A.' Peebles, Chief
b , f-/.7 ,f7
Date Signed
Projects Section 3A
1
l
i
,
I
8805020106 880420
PDR ADOCK 05000413
Q DCD
~
4
4
. .
TABLE OF CONTENTS
Page
I. Introduction - Formation and Initiation of AIT
A. Background 1
B. Formation of Augmented Inspection Team (AIT) 1
C. AIT Charter - Initiation of Inspection 1
D. Persons Contacted 3
E. Design Descriptions 3
1. Nuclear Service Water (RN) System 3
2. Auxiliary Feedwater (CA) System 7
3. Condensate Storage (CS) System 8
II. Description of Events
-
Overview of Event for Catawba Unit 2
1. Event Description 9
2. Detailed Sequence of Events 12
III. Equipment Status and Evaluation
A. Auxiliary Feedwater Suction Swapover to
Nuclear Service Water System 1
1. Control Logic for CS/RN Swapover 14
2. Equipment As Found 16
3. Suction Swapover Event Analysis 19
4. Dynamic Test 22
5. Dynamic Testing Conclusions 24
6. Conclusions 24
7. Corrective Actions 25
B. Investigation of Clam Fouling of RN Supply
1. History of Licensee Programs for RN Fouling 25
2. Flushes and Inspection 28
3. Evaluation of Short Term Corrective Actions 34
4. Long Term Corrective Actions 36
IV. Regulatory Requirements 37
V. Conclusions 39
VI. Exit 40
Attachment 1 - Three (3) Graphs of Dynamic Testing
Attachment 2 - Two (2) Detail figures of CA Flow Control Valve (2CA56)
. Attachment 3 - Three (3) Letters of Concurrence
! Attachment 4 - Three (3) Training Diagrams of CA
i
. _ . .__. - _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ - _ _ _ _ _ _ _ _
6.
4
s. :
, ,
REPORT DETAILS
I. INTRODUCTION - FORMATION AND INITIATION OF AIT
A. Background
Catawba Units 1 'and 2 are four loop Westinghouse Pressurized
Water Reactors rated at 1145 megawatts electric. 'They have ice
condenser containments and are located about 25 miles south of
Unit 1 was licensed January 17,
~
Charlotte, North Carolina.
1985, and Unit 2 in May 25, 1986.
On March 9, 1988 at 10:15 PM, the licensee notified the NRC
headquarters duty officer of the following event:
Catawba Unit 2 reactor tripped from 21% power on low-low
steam generator (S/G) level on "B" S/G. This occurred when
the feedwater regulating valve on "B" S/G (2CF37) failed
open (cause unknown and being investigated) when being
placed in auto. This caused a feedwater swing which
resulted in "C" and "D" S/G to be fed up giving a P14
signal (high-high S/G 1evel) on the "0" S/G. On the P14
signal, a feedwater isolation occurred which resulted in
all S/Gs levels decreasing and the "B" S/G reaching the
low-low level giving a reactor trip signal. Both auxiliary
feedwater pumps auto started but for an unknown reason the
"A" auxiliary feedwater (CA) pump suction swapped over to
the nuclear service water (RN) system.
B. Formation of AIT
On the morning of March 10, 1988, the Regional Administrator,
after briefing by regional staff and consultation with senior
NRC management, directed the dispatch of an AIT.
C. AIT and Charter
The charter for the AIT was prepared on March 10, 1988, and the
AIT members began arriving at the Catawba site that morning.
The licensee shutdown the other unit at 10:50 a.m. due to their
concerns of CA operability. An entrance meeting and briefing by
the site staff and plant manager was held with the AIT at
5:00 p.m.
A Confirmation of Action letter (CAL) was issued on March 11,
1988. This CAL documented the understanding between the
licensee and NRC that Region II would be in agreement with
investigation results prior to startup of either unit.
.
_.
!q.
4
. .
The charter for the AIT specified that the following tasks be
completed.
1. Develop and validate a detailed sequence of events
associated with the March 9, 1988 Reactor - Trip and
subsequent swapover of the suction to the Auxiliary
Feedwater Pumps to the Nuclear Service Water System and
subsequent degraded auxiliary feedwater flow.
'2. Evaluate the significance of the event with regard to
radiological consequences, safety system performance,
safety significance, and plant proximity to safety limits
as defined in the Technical Specifications.
3. Evaluate the accuracy, timeliness, and effectiveness with
t.hich information on this event was reported to the NRC.
4. For nach equipment malfunction, to the extent practical,
determine:
a. Root cause;
b. If the equipment was known to be deficient prior to
the event;
c. If equipment history would indicate that the equipment
had either been historically unreliable or if
maintenance or modifications had been recently
performed;
l d. Any equipment vendor involvement prior to or after the
event;
- e. Pre-event status of surveillance, testing and/or
- preventive maintenance; and
f. The extent to which the equipment was covered by
existing corrective action programs and the
irrplication of the failures with respect to program
effectiveness.
5. Evaluate the licensee's action taken to verify equipment
operability on the operating unit.
6. Identify any human factors / procedural deficiencies related
to the event.
,
f
y -_ - -_ _ - - - . _ - . - _ _ -
'
5
4'.
I
19-
. -,
7. Through o'perator and technician interviews, determine if
any of 7the following played a significant role in the
event: plant. material condition; the quality _ of
maintenance; or the responsiveness of engineering to
identified problems.
8. Provide la Preliminary Notification upon initiation of the
-inspection and an update'<on the- conclusion of the
'
inspection..
9. Prepare a special inspection report documenting the results
of the above activities within 30 days of the start of the
inspection.
D. Persons. Contacted
Licensee Employees at Exit Meeting
T. B. Owen, Manager, Catawba Nuclear Station (CNS)
C.-L. Hartzell, Compliance Engineer, CNS
M. A. Cote, Licensing Specialist, CNS
C. E. Muse, Unit 2 Coordinator, CNS
R. F. Wardell, Technical Services Superintendent, CNS
W. R. McCollum, Station Services Superintendent, CNS
G. T. Smith, Maintenance Superintendent, CNS
Othe'r Licensee Employees
E. Fritz
M. Carwile
J. Kammer
M. Anderson
J. Caldwell
Z. Taylor
Other licensee employees were contacted.
E. Design Descriptions
1. Nuclear Service Water System (FSAR Section 9.2.1)
The Nuclear Service Water System (RN) provides essential
auxiliary support functions to Engineered Safety Features
of the i.sation. The system is designed to supply cooling
water to various heat loads in both the safety and
non-safety portions of each unit. Provisions are made to
ensure a continuous flow of cooling water to those systems
and components necessary for plant safety during normal
operation and under accident conditions. Sufficient
redundancy of piping and components is provided to ensure
that cooling is maintained to essential loads at all times.
.
La.
4
. .
Functionally, the system consists of four sections which,
when put together in series, serve to assure a supply of
river water to various station heat loads and return the
heated effluent back to its proper heat sink.
In order of flow, these are:
a. Source and intake section
b. RN Pumphouse section
c. Station heat exchanger section
d. Main discharge section
Source and Intake Section
Two bodies of water serve as the ultimate heat sink for the
components cooled by the RN System. Lake Wylie is the
normal source of nuclear service water. A single transport
line conveys water from a Class 1 seismically designed
intake structure at the bottom of the lake to both the A
and B pits of the Nuclear Service Water Pumphouse serving
the RN pumps in operation. Isolation of each line is
assured by two valves in series and fitted with electric
motor operators powered from separate power supplies.
Should Lake Wylie be lost due to a seismic event in excess
of the design of Wylie Dam; the Standby Nuclear Service
Water Pond (SNSWP), formed by the Class 1 seismically
designed SNSWP Dam, contains sufficient water to bring the
station safely to a cold shutdown condition following a
single loss of coolant accident. The SNSWP has an intake
structure designed to Class 1 seismic requirements, with
two Class 1 seismic, redundant lines to transport water
independently to each pit in the RN Pumphouse. Each line
is secured by a single motor operated valve. Automatically
upon loss of Lake Wylie the isolation valves are closed and
the SNSWP valves are opened to both pit A and pit B.
RN Pumphouse Section
The RN Pumphouse is a Class 1 seismically designed
structure that contains two separate pits from which two
independent and redundant channels of RN pumps take
suction. Each pit can be supplied from both the normal
source and also the assured source of water. Either pit is
capable of passing the flow needed for a simultaneous unit
LOCA and unit cooldown. Flow spreaders in front of all the
intake pipe entrances prevent vortices and flow
irregularities while removable lattice screens protect the
RN pumps from solid objects,
t
, _. - - _ _ _ _ _ _ _ _ _ - __ . _ _ _ . _ . _ _ _ - _ _ ___ _ - _ _ _ _ _ .__ __-_-_. _ _ _ _ _ _ _ _ - _ _ _ _ _ _
..
'O.
. .
pumps IA and 2A take suction from pit A and discharge
through RN strainers 1A and 2A respectively. The outlet
piping of the 1A and 2A RN strainers then join back
together to form the channel A Supply line to channel -A
components in both units.
RN pumps 18 and 28 are physically separated from RN pumps
IA and 2A by a concrete wall, and take suction from pit B,
discharging through RN strainers 1B and 2B respectively.
The . outlet piping of strainers 18 and 2B join together to
form the channel B supply line to channel B components in
both units.
The supply and return headers are arranged and fitted with
isolation valves such that a critical crack in either
header can be isolated and will not jeopardize the safety
functions of this system or flood out other safety related
,
equipment. The operation of any two pumps on either or
both supply lines is sufficient to supply all cooling water
requirements for the two unit plant for unit startup,
cooldown, refueling, or post-accident operation. However
additional pumps are normally started for unit startup and
cooldown and two pumps per unit operate during the
hypothetical combined accident and loss of normal power if
both diesel generators are'in operation. In an accident
the safety injection signal automatically starts both RN
pumps in each unit, thus providing full redundancy.
Heat Exchanger Section
Nuclear Service Water supplied by the RN pumps is used in
both units to supply essential and non-essential water
needs or as an assured source of water for other
safety-related systems.
Essential components are those necessary for safe shutdown
of the unit, and must be redundant to meet single failure
criteria. Nonessential components, are not necessary for
safe shutdown of the unit, and are not redundant. Each
unit has two trains of essential heat exchangers designated
A and B, and one train of nonessential heat exchangers
supplied from either A or B and isolated on Engineered
Safety Features actuation.
The following components or services are supplied by each
essential header of the RN System. Some components are
normally in operation, some are automatically supplied upon
ESF actuation, and others are used when needed.
4
4
. .
a. RN Pump Motor Cooler
b. RN Strainer Backflush
c. RN Pump Bearing Lube Injection Water
d. RN Pump Motor Upper Bearing Oil Cooler
e. Diesel Generator Engine Jacket Water Cooler
f. Diesel Generator Building Essential Fire Water
g. Diesel Generator Engine Starting Air Aftercooler
h. Component Cooling Heat Exchanger
1. Assured Auxiliary Feedwater Supply
j. Assured Fuel Pool Makeup
k. Assured Component Cooling (KC) System Makeup
1. Containment Spray Heat Exchanger
m. The Control Room Area Chiller Condenser (Condensers A
and B are shared between units, so they are fed by
Unit 1 Essential Headers only).
n. Auxiliary Shutdown Panel Air Conditioning Unit.
o. Assured Containment Penetration Valve Injection Water
(RW) System Makeup
From each essential RN header, a six inch branch line
provides the backup water supply for Auxiliary Feedwater.
This six inch line travels for approximately 250 feet
before the first isolation valve. Just upstream of the
isolation valve, a line branches off to supply
approximately 10 GPM continucus flow to each auxiliary
shutdown panel air conditioning unit condenser.
No provision is made for prevention of long-term corrosion
in the RN System. Allowances for such corrosion were made
by increasing the wall thickness of the pump pressure
boundary, piping, and the heat exchanger shells and tubes
in accordance with the applicable codes. Larger pipe sizes
than necessary were used for pump adequacy considering
scaling.
Asiatic clam control is achieved by a series of design
features and operating procedures. Intake structures take
suction elevated off the bottom of the lake and SNSWP and
at a velocity well below that required by the Environmental
Protection Agency for wildlife protection, so large clams
are not drawn into the system. A bar screen with openings
4 in. x 4 in, keeps debris from entering the intake lines
and a lattice screen with 1 in x 1 in, openings separates
the forebay of the RN Pumphouse from the RN pump suction
bay, providing a second level of defense. The water
discharging from each RN pump passes through an RN strainer
with 1/32 inch openings to strain out dirt and sand
particles that could clog control valves with cavitrol trim
located throughout the RN System. These screens and
strainers will prevent all but the smallest clam larvae
from entering tha RN piping.
_ - _ _ _ _ _ - _ _ _ - _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ . _ _ - _ _ _ _ _ . . _ _ _ _ ___ _____________ ___________ _________-_____-
_ _ _ _ _ - _ _ _ _ _ - _ _ _ _ _ _ _ _ _ - _ _ _ - - - _ _ _ _
-o-
.
4*) . .
7
It is understood that Asiatic clam larvae do not
permanently attach to pipe walls and grow, but nest in
stagnant places such as valved off pipes and idle heat
exchangers and operating heat exchanger heads in front of
the tube sheet, performance monitoring programs to verify
adequate flow, and visual inspection of the intake piping
and inlet heat exchanger heads during maintenance provide
early detection of any clam infestation of raw water
systems.
2. Auxiliary Feedwater System (FSAR Section 10.4.9)
The Auxiliary Feedwater System (CA) assures . sufficient
feedwater supply to the steam- generators (S/G), in the
event of loss of the Condensate /Feedwater System, to remove
energy stored in the core and primary . coolant. The two
units are provided with separate CA Systems.
The CA System is designed to start automatically in the
event of loss of offsite electrical power, trip of both
main feedwater pumps, safety injection signal, or low-low
S/G water level; any of which may result in, coincide with,
or be caused by a reactor trip. The CA System will supply
sufficient feedwater to maintain the reactor at hot standby
for two hours followed by cooldown of the Reactor Coolant
System (NC) to the temperature at which the Residual Heat
Remocal System (ND) may be operated.
Three CA pumps are provided, powered from separate and
diverse power sources. Two full capacity motor driven
(MDCA) pumps are powered from two separate trains of
emergency on-site electrical power, each normally supplying
feedwater to two steam generators. One full capacity
turbine driven (TDCA) pump, supplying feedwater to two
steam generators, is driven from steam contained in either
of the two steam generators. A minimum of 225,000 gallons
feedwater supply is required for the design basis hot
standby followed by normal cooldown to conditions at which
the ND System may be operated.
Standards for nuclear safety related systems are met for
the CA System except for the condensate quality feedwater
sources.
The normal lineup is for train A MDCA pump 1(2)A to supply
A and B steam generators, for train B MDCA pump 1(2)B to
supply C and D steam generators and for the TDCA pump to
supply B and C steam generators,
.
4
- *
'
8
.
-
The MDCA pumps will automatically start and provide the
minimum required feedwater flow within one minute following
any of these conditions:
a) Two out of four low-low level alarms in any one of the
i four steam generators,
b) Loss of both main feedwater pumps,
c) Initiation of the safety injection signal.
d) Loss of station normal auxiliary electrical power.
The TDCA pump will automatically start and provide the
minimum required feedwater flow within one minute following
either of these conditions:
a) Two out of four low low level alarms in any two of the
four steam generators.
'
b) Loss of station normal auxiliary electric power.
Driving steam for the TDCA pump is provided from either. the
steam generator B or C main steam lines upstream of the
main steam lines containment isolation valves and is
discharged to the atmosphere from the turbine.
The Condensate Storage System provides a readily available
source of deaerated condensate for makeup to the condenser
and is the preferred source of auxiliary feedwater for
makeup to the steam generators. It also serves to collect
and store miscellaneous system drains.
Makeup to the Condensate Storage System is to the upper
surge tank (UST) dome from the Makeup Demineralized Water
System. The upper surge tank dome drains to the two upper
surge tanks. Makeup to the condenser is supplied by
gravity flow from the UST. The UST also provides sealing
water for various equipment, and the supply for the
auxiliary electric boiler feedwater pumps. Overflow from
the UST is returned to the Condensate Storage Tank (CST)
through a 27 foot loop seal which prevents the introduction
of air into the upper surge tanks. Valve ICM363 provides a
means for filling and makeup to the loop seal. The CST
receives the drains from various equipment and holds these
drains until they are transferred to the UST dome by the
two full capacity CST pumps. The CST operates at
atmospheric pressure and is vented to the roof. The
overflow line from the CST has a 3 foot loop seal to
prevent steam from the hot drains discharging into the tank
from entering the building.
--- . _ _ _ _ . _ _ _ _ - _ - -- _-.
- '
0' f
i
I
.. .
9 i
l :
) The preferred source of clean' water supply for the l
'
auxiliary feedwater pumps is provided by the main condenser i
hotwell (170,000 gallons) and the two UST (85,000 gallons ;
total) on each unit and the shared Auxiliary Feedwater :
Condensate Storage Tank (CACST) (42,500 gallons). Relative j
pressure differentials considering normal vacuum in the
hotwell and UST result in the supply to the CA pumps !
initially from the CACST -then th9 UST. The hotwell may
then be pumped to the UST or vacuum relieved and supplied
'
i
directly to the CA pumps. The Condensate Storage System ,
i tanks are not safety related, since the. assured source of ;
'
water for the auxiliary feedwater pumps is provided from !
the Nuclear Service Water System. j
Sufficient instrumentation is provided to monitor system [
performance. Alarms are provided in the control room for !
high and low UST level, high and low CST level, low hot
well level and low CACST tank level.
All of the preferred sources of condensate quality water- ;
are normally aligned to the CA pump suctions. The l
condensate reserve for-each unit is maintained among the ;
i following sources: l
I
Source Max. Capacity .j
t
a. Auxiliary Feedwater 42,500 gallons / station i
1 Condensate Storage Tank (Shared by both units) !
I
b. Upper Surge Tanks 85,000 gallons / unit !
(two 42,500 gallon tanks !
per unit) j
-
>
! c. Condenser Hotwell 170,000 gallons / unit l
- . (Normal operating level) !
- i
!
II. DESCRIPTION OF EVENT
.
.
Overview of Event for Catawba Unit 2 l
i
i 1. Event Description !
! At 6:25 pm on March 9,1988, Catawba Unit 2 was operating 7
at 20*4 reactor power carrying load of 150 MWE and in the -
.
process of starting up from the unit's first refueling. An
,
, i
operator was transferring from the feedwater (CF) bypass i
valve to the main feedwater regulating valve (2CF37) on l
! . steam generator (S/G) B. As the valve (2CF37) was placed [
in automatic, it failed full open which caused the ;
i
b
i
'
I
l
. - . _ . , , - - - , .- - -,. ,
, . - - - - - - - _ _ _
':
,
. .
10
operating main feedwater pump speed to start oscillating
thus resulting in a major swing in flow throughout the feed
system. Another operator assumed manual control of the
pump and attempted to bring it under control. The operator
on the B main feedwater regulating valve assumed manual
control and closed it down to prevent over feeding S/G B.
Level in S/G's C and 0 meanwhile increased until two of
four channels on S/G 0 reached High/High S/G level. This
resulted in a main turbine trip, a feed water isolatior,
and a main feedwater pump trip.
Both motor driven auxiliary feedwater (MDCA) pumps auto
started immediately upon loss of the main feedwater pump.
An apparent CA pump low suction pressure resulted in the
swap of the A train MDCA pump suction to the RN system
which serves as the safety related assured source (valves
2RN 250A and 2CA 15A opened). Details concerning the swap
to RN are delineated in Section III. A of this report.
Level in S/G's A and B were meanwhile dropping and a
reactor trip occurred due to Low / Low Level in S/G A. The
turbine driven auxiliary feedwater (TOCA) pump was
initiated when S/G B also reached Low / Low level which
satisfied the logic of two of four S/G's at Low / Low level.
The operators secured the TDCA pump approximately one
minute after it auto started and manually throttled CA flow
to stabilize S/G levels.
Approximately 13 minutes into the event, the control room
operators reviewed the panels to determine any abnormal
indications. It was at this time that one operator saw an
annunciator illuminated which indicated that a CA suction
swapover had occurred and another operator detected (on the
control panel) that valve 2RN 250A was open. The operators
immediately closed 2RN 250A, and verified suction flow path
from the UST. It should be noted that neither operator
detected that valve 2CA 15A had opened. (Valve 2RN 250A
and valve 2RN 15A are in series and both must open for RN
to be surplied to MDCA pump 2A.)
Approximately 20 minutes into the event, with auxiliary
feedwater maintaining steam generator levels it was
detected that the level in the 8 generator was starting to
decrease slightly. The operator opened the CA flow control
valve supplying the B S/G (2CA 56) but noticed that the
valve when full open would pass only 100 gpm, instead of
normal flow of 300 gpm. This was the first indicstion of
flow degradation even though the operators at this time
still had not detected that valve 2CA 15A had opened in
response to the apparent low suction pressure signal.
_ _ - _ _ _ - - . _ _ _ _ _ _ _ - _ _ .
.
'4
I
. .
11
.
i
At approximately 7:00 pm, some 35 minutes into the event, "
during shift turnover, the oncoming shift supervisor
detected. that valve 2CA 15A was open. It was not until- >
this time that the on duty operating shift knew that
2CA 15A- had repositioned. Furthermore, it was not until
this time that the on duty Unit Supervisor and Shif t :
Supervisor became aware that the unit had experienced a CA !
suction problem. This is indicative of a communications ,
problem on the part of the on duty operations crew. "
Auxiliary feedwater continued to supply the generators _ !
until about 8:45 pm when the operators were able to restart i
the main feedwater pump, and place the CA system in
standby,
t
After it was detected that valve 2CA 15A had opened,
actions were initiated to disassemble and inspect CA flow '
control valves 2CA 56 and 2CA 60, the valves which had
passed RN water to S/G's A and B. By 10:30 am March 10, it
was determined that the valves were clogged with clam
shells.
By 11:50 a.m. March 10, it had also been concluded that .
continued safe operation of unit 1, which was at 100% i
reactor power, could not be assured due to the possibility
of clam fouling of the CA system. At 12:50 p.m. the .l
licensee began shutting unit I down to mode 4. At 4:45 ;
p.m., unit 2 was taken into mode 4. '
The licensee formulated a program of flushes and ,
inspt:ctions to confirm that suspect RN lines supplying [
assured sources of cooling to safety related equipment were :
clear of fouling. Details of these programs are delineated *
in Section III.B of this report.
!
The licensee also initiated a test and inspection program
to verify the correct operation of the instrumentation, j
logic and equipment associated with the CA suction
'
swapover, as well as CA pump verification tests and CA ,
system valve function. Details of the testing performed ,
are described in Section III.A. of this report.
Subsequent to the completion of the aforementioned flushes,
and inspections on March 11, 1988, unit I restarted af ter ;
conferences with the AIT, the Region and NRR. Unit 2 was ,
allowed to restart on March 18, 1988 after resolving the r
issues of fouling and CA suction swapover.
,
f
. _ _ _ _ - _ _ _ _ _ _ _ _ - - _ _ _ _ _ _ _ _ _ _ _ _ - _
o
4
. .
12
2. Detailed Sequence of Eveny
The sequence of events was developed from discussions with
operations personnel, review of computer alarm typer data
recorded during the event and review of plant logs and
investigation packages.
March 9, 1988
6:25:00 PM - Unit operating at 21*e power carrying load of
150 4'E.
6:25:37 - Received Turbine Trip. Feedwater isolation
and Feedwater Pump trip on High/High S/G
1evel in the D S/G when main feedwater
regulating valve 2CF37 failed open when
placed in automatic,
6:25:37 -
Both MDCA pumps started on the loss of the
running main feedpump.
6:25:44 -
Low CA pump suction pressure resulted in
swap of suction to Nuclear Service Water,
the assured source.
6:25:49 - Levels in A and B S/G's decreasing; level l
in S/G A reaches low / low level in 2 of 4
channels resulting in reactor trip.
6:26:09 -
Level in 8 S/G reaches low / low level in 2
of 4 channels - results in autostart of TD
CA pump.
6:27:20 -
TOCA pump secured.
-
Operators throttle CA flow to S/G's to
stabilize cooldown rate.
6:38:55 -
Operators isolate RN from CA pump suction by
closing valve 2RN 250A.
6:45 approx. -
It was noticed that CA flow to the A and B
S/G's had degraded over time. Initial flow
was normal (300 gpm/SG) but had decreased to
approximately 100 gpm to the A S/G and 200
gpm to the B S/G. Operator notes that with
valve 2CA 56 (CA flow control valve to the A
S/G) full open, valve will pass only 100
gpm.
10:15 PM Licensee notified NRC Duty Officer of Unit 2
___ _ .. . __-. _ _ _ _ .
' * ,
.-
. 1
e i
13
!
i
March 10, 1988 !
!
6:00 AM - Licensee disassembles and-inspects CA flow
10:00 AM control valves 2CA56 and 2CA60 which supply
the A & B S/G's. . Upon disassembly, valves i
were found to be clogged with clam shells, j
9:00 AM Licensee discusses event'with Region II. }
-
Licensee evaluating cause of CA swap to RN. .
11:50 AM' - Licensee determines continued safe operation !
of Unit 1 can not be assured due to the '
possibility of clams fouling the CA system, I*
>
should swapover to . RN become necessary.
Unit shutdown began. !
12:30 PM Licensee notified NRC Outy Officer of the :
shutdown of both units. i
12:53 PM Licensee notified NRC Duty Officer of !
Unusual Event declared on Unit 1 as a safety i
system i
(CA) was degraded at power and shutdown of I
both units was commencing. [
11:15 PM Unusual event terminated. Unit 1 in mode 4,
March 11, 1988 .
l
- Licensee initiater program to flush stagnant i
portions of RN piping which serve as back up i
supply to safety systems, Flushing program i
includes radiography and boroscopic exami- -i
nations to confirm flush adequacy. [
- Licensee initiates testing on both units CA i
- to RN swap logic and equipment to determine :
if calibrated, operable and cause for swap t
on Unit 2. l
-
CAL Issued f
11:00 PM - All testin0 and flushing complete on Unit 1, !
unit receives NRC concurrence for restart. I
f
-
Unit 2 flushing and testing continuing. l
!
March 13, 1983 I
9:18 PM -
Unit 1 enters Mode 2 i
i
March 12-18, 1988
- Flushes, inspections and tests continue, j
l March 18 -
Unit 2 allowed to restart commensurate with
j CAL. Licensee agreed to utilize
compensatory measures relative to CACST l
valve 2CA 6. }
I
'
. _ - _ _ _ _ - _ _ _ _ _ . _ - - .__
v .,. -
r
o
r
,
,_ , . . f
14 4
t
!
III. EQUIPMENT STATUS AND EVALUATION j
!
- - A. Auxiliary Feedwater Suction Swapover to Nuclear Service Water [
System
'
This section discusses the Catawba Unit 2 CA suction swapover l
.
from the normal sources to the safety-related assured source of r
RN. To cover the March 9,1988, event in detail the following .r
sections are included:
. 1. Control icgic '
t
,
2. Equipment as found
3. Conclusion
i
1. Control Logic for CS/RN Swapover [
l f
"
As the CA System serves a vital safety related function r
during all postulated occurrences two trains supplied by a
safety grade, seismically designed water source must be ,
assured at the pump suctions to assure pump operability and l
function. The Standby Nuclear Service Water Pond, with a i
maximum capacity of 2.74 x 10' gallons total for the !
station, serves as the ultimate long-term safety related I
source of water for the CA System. To maintain steam l
generator water chemistry, especially for events which t
4 require fast response recovery such as; blackout, loss of f
i'
"
normal feedwater, or main steam system malfunction, the CA l
pumps are normally aligned to condensate quality water, i
The sources of condensate quality water exist as non- ;
seismic grade sources in the Turbine Building. To prevent :
inadvertent injection of out of chemistry nuclear service [
water to the steam generators a reliable means of detecting l
1 loss of condensate source and automatic transfer of the !
pump suctions to the nuclear service water sourco is [
t employed. Such detection and transfer controls are ;
automatic since minimum CA System flow must be established !
within one minute from the initiating event. The automatic #
detection and transfer controls will detect and transfer .
the pump suctions to nuclear service water upon detection i
i of any of the listed postulated failures of the non-seismic
I condensate supplies:
(a) Deeletion of all condensate sources.
Losa of source due to pipe break.
,
(b)
(c) Partial or complete loss of source due to air leakage
2
into the system from a pipe crack, or failure to t
'
l
isolate a depleted source.
4
(d) Partial loss of source due to steam void formation in j
the suction piping caused by excessive friction loss i
associated with a high flow rate, failure or spurious !
i operation of a valve causing partial closure, or [
bending or partial obstruction in the pipe. l
l
}
,
, ,= - .-
'O- '
t
'
f*e,.
5 :.
15
o
'
~
. The detection scheme incorporates three trains of three
differential pressure switches located in the Auxiliary
-Building in a vertical leg of the common condensate supply
. pipe to all three CA - ~ pump s . Two trains of pressure
switches . serve the two safety grade RN trains, and the
-remaining train of pressure switches serves 'the condenser
circulating (RC) system supply. RC provides a backup to
RN for further CA suction supply. One of three pressure ,
y switches activating gives a low pressure alarm in the
'~
control room. Upon two out of three indications of low
suction pressure from any train, the transfer logic will be
activated for that train. The instrumentation and controls
for the RN trains meet the standards for nuclear safety
related systems, including requirements for redundancy and
separation. If the station normal auxiliary electrical
power is available during the initiating occurrence, a
maximum 30,000 gallons additional condensate supply is
available from the condensate storage tank. If the two
makeup demineralizers are available, a maximum condensate
supply of 950 GPM is available for the short term or 475
GPM for an indefinite period. Additional condensate . may
also be provided from condensate sources associated with
the other unit, if these sources are available, operable,
and a loss of normal station auxiliary electrical power has
not occurred.
The logic for the automatic swapover of the CA system to
take suction from the nuclear service water (P.N) system is
based upon the following conditions:
1. CA pump running
2. Loss of condensate source (i.e. 2 out of 3 low
suction pressures at pressure switches)
. The Technical Specifications assume an overall
time for swapover to occur of 15 seconds with 10
seconds allowed for valve movement.
Applied to the above logic is a time delay (3-5 seconds)
for the low suction pressure. The purpose of the time
delay relay is to allow for a momentary low suction
pressure during a pump start transient. f-
The automatic switchover to RN takes place only if the CA
system has been initiated by a CA auto start signal and the
CA pump suction pressure is low as sensed by 2/3 pressure
switches. Low pressure in one of three A or B train
pressure switches will alarm the Onerator Aid Computer
(OAC). Either train A or B RN st.ction sources can align to
the turbine driven CA pump if all of the following has
occurred within the proscribed sequence:
(1) The TDCA pump has received an auto start signal.
.r . . . _ . . - . - . . _
, .
-
...~ . . -
3
p <. +
i' -- .
,;
,
n
- .- .
'
16
.
>
(2) The pump turbine steam supply: valve 'SA2 or SAS has .
opened.
(3) The-pump turbine trip and throttle valve is open.
(4) A 3-5 second time delay -times out .pr.ior to providing
the'run indication to the swapover logic of TDCA pump.
This time delay relay is in addition to-the 3-5 second
'
time delay intended to prevent _ swapover on a pump
start transient.
(5) Two out of three low section pressure indications from
the train A or B pressure switches and their 3-5
second time delay relay has timed out.
CA purnp -suction following swapover to RN is through the
following paths:
Train A is from the 24 inch A troin RN header to the ,
6 inch RN line through valve RN 250A then either
through valve CA 15A to the A train MDCA pump or
through valve CA 116A to the TDCA pump. ;
Train B is from the 24 inch B :cain RN header to the
6 inch RN line through valve . RN 3103 then either
through valve CA 18B to the B traia MDCA pump or
through valve CA 85B to the TOCA pump. ,
The following section lists all the as found setpoints for
. trains 1A, 1B, 2A, and 28,
2. Equipment as found -
(a) pressure Switches
(1) Unit 1 A Train
Required setting: 10.5 .15 psig
.- Pressure Switch As Found Setpoint
1 CAPS 5220 .'.0. 8 p si g
l.
1 CAPS 5221 10.44 psig
1 CAPS 5222 9.92 psig
Technical Specification (TS) trip setpoint
i 3 10.5 psig
l
TS Allowable Value 1 9.5 psig
!
!
!
I
..- . - - . - - - _ - - . _ - . - . , , . . . - ,. - , .-.
, .
.
' "
. : ,.. .
'
..:
1
S '.g' /
.q
.
m -
37
-
. .
(. ", J
' '
(2) Unit 1 B Train
e
.Raquired: setting: 6.0 .15 psig
is Pressure Switch- As Found Setpoint~
1 CAPS 5230 5.95 psig-
'
-1 CAPS 5231 .5.90 psig
-1 CAPS 5232 6.10 psig .
TS trip setpoint 6.2 3 psig
,
TS Allowable Value t 5.2 psig
l
(3) Unit 2 A Train
Required setting: 10.St.15 psig
Pressure Switch As Found Setpoint
,
2 CAPS 5220 11.0 psig
2 CAPS 5221 11.2 psig
2 CAPS 5222 10.6 psig
TS trip setpoint 310.5 psig
. TS Allowable Value 1 9.5 psig
(4) Unit 2 B Train
Pressure Switch As Found Setpoint
2 CAPS 5230 6.85 psig
2 CAPS 5231 6.05 psig
2 CAPS 5232 6.05 psig.
l
TS trip setpoint t 6.0 psig
TS Allowable Valve 3 5.0 psig
! It was noted that the licensee interprets the Trip
I
Setpoint, as specified in Table 3.3-4 of TS, as a
nominal value and allows a band of .15 psig. The TS
indicates that the trip setpoint is a minimum value as
opposed to a nominal value, although TS basis
discusses it as a nominal value. The licensee's
interpretation appears to be consistent with the
Westinghouse setpoint methodology. This was confirmed
with R. Giardina of NRR, OTSB.
- .__ - _ _ _ _ - __
L. . . .
.
f4!
f.
v . .,
18
r (b) Time Delay
,
o
i: The Unit 1 A Train time delay was found to be 4.27
b seconds and for Unit 1 B Train it was 4.39 f seconds.
The Unit 2 A Train time delay was found to .tme .4.15
[ ,
L . . seconds and for Unit 2 B Train it was 4.7 seconds.
These values . were found acceptable because they
conform to the manufacturer's recommendations of being
in the range of 3-5 seconds.
(c) Pressure Switch Root Valves
Pressure switches 2 CAPS 5220, 5221, 5222 have common
impulse lines and are isolated by one root valve.
. Pressure switches 2 CAPS 5230, 5231, 5232 have common
impulse lines and are isolated by one root valve. The
root valves on Unit 2 A and 2 B Trains were found to be
partially blocked by magnetite. The blockage may have
contributed to the fact that the alarm in the control
room had a 51 second duration before it cleared.
(d) Valve 2CA6
Valve 2CA6 is the supply of condensate grade water to
the CA pump from the CACST and is powered from the
emergency bus. The valve is normally open and for 3
days af ter the event the licensee believed the valve
had been open. However, after completing their
'
interviews of the operators, the licensee determined
that 2CA6 had been closed at the start of the event.
The same valve on Unit I had been closed at various
times over the past year to prevent drainage of the
CACST through the CA system. The CACST is a common
tank to both units and was receiving makeup from Unit
1 at the time. The licensee had been attempting to
identify the causes for CACST level decreases.
The losses of CACST ievel were not fully understood
by the licensee but were believed to be a result of
check valve backleakage to the hotwell or a design
feature.
The operations r.taff had decided to close 2CA6 on
Unit 2 to prevent migiation of water between Units 1
and 2 and to prevent draining the CACST. The CACST
normally supplies the highest head and thus shutting
2CA6 apparently removed this added margin and
contributed to the low suction pressure event.
- _ _ - - _ .
- _ _ _ _ _ _ _ _ _ _ _ .
-_-- __
,. . . .
. ._
\
-
. .:
s
. . .
<
19
(e).'Swapover Valves
All of the major Unit 2 train A and B Swapover valves-
had been periodically stroke tested and were tested
and found operable following the event.
3. Suction Swapover Event Analysis
Normally all three CA pumps are aligned to non-safety
grade, condensate -quality sources. The safety-related RN
supply-is not normally aligned to the CA pump suction, but.
is automatically aligned when low suction pressure is
detected.
The following evaluation of the March 9,1988 sequence of
events is based on analysis of strip charts, computer
l- alarm printouts, calibration of components and interviews
'
of personnel. This is the AIT evaluation which is
consistent with that of the licensee's evaluation.
Both MDCA pumps 2A and 2B started after the main feedwater
pump trip. The OAC received a low pressure alarm from both
~
the train A and train B pressure switches. Valves 2RN250A
and 2CA15A opened, aligning the RN train A suction source
with MDCA pump 2A. MDCA pump 2B was still aligned to the
normal suction soure.e. MDCA pump 2A was operating while
taking suction fr'm its assured RN source, and
simultaneously pump c3 was operating while taking suction
from the normal non-safety grade sources. About 32 seconds
af ter the MDCA pumps star..ed, valves SA2 and SAS opened,
and the TDCA pump began to ramp up to 3650 rpm and pump
auxiliary feedwater to stear.' generators B and C. The pump
had received a CA pump start signal on detection of low low
level in two of four steam generators. Nineteen seconds
after valves SA2 and SAS opened, the low suction condition
in train A and B cleared.
Based on the actuation sequence of the CA pump suction
pressure switches, it appears that when the CA pumps
started, the suction pressure dropped below the actuation
pressure of the decreasing setpoint on at least two of the
train A pressure switches, (2CAPSS220, 5221, and 5222) and
below the decreasing setpoint of one of the train B
switches, (2 CAPS 5230, 5231, and 5232). No information is
available that indicates condenser circulating water (RC)
pressure switches (2 CAPS 5240, 5241, and 5242) actuated.
._ - .
_ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _
_ _ _ _ - - - _ _ _ _ _ _ _ _ _ _
- .
..
?f.
. .
20
The pressure appears to have stayed below the reset
pressure of the two train A switches for at least four
seconds, and . initiated swapover. to RN as a source for MDCA
purrp 2A. . After about 51 seconds, the pressure apparently
rose above the reset point of the train B sw:tch and all
the train A switches.
As two of the three train A pressure switches had actuated
and a partial train A swapover to - RN occurred, it was
questioned why valvo 2CA116A did not open during the
March 9 event and Clow RN to supply the TDCA pump.
After the initial s, over occurred, one of the pressure
switches could have reset between five and thirty-two
seconds into the event, aided by the pressure increase from
the RN supply, and prevented valve 2CA116A from opening.
The limit switch indicating valve 2CA116A closed on the
computer was.not functioning, so no positive indication of
valve position was available. However, since: 1) the TDCA
pump auto start signal occurred approximately 32 seconds
af ter the MDCA pump auto start signal; 2) a time delay
occurs in opening 2SA2 or 2SA5; and 3) a 3-5 second time
delay relay times out before the low suction pressure 3-5
second time delay relay starts the 2/3 low suction pressure
logic; the logic could have partially cleared at some
point in this sequence and demand for 2CA116A to open was
never actuated.
The licensee initially suspected that valve 2CA116A had
partially opened because chemistry results of the 2C S/G
indicated high levels of cation conductivity. Several days
later it was determined that these samples were in error.
The sample line from C S/G became clogged af ter the trip
and the chemistry technicians were actually sampling A or B
S/G. The chemistry of steam generators C and D was found
to be actually within specification indicating that valve
2CA116A did not open. Further inspections were conducted
on valve 2CA116A; it passed a stroke test following the
event and it was disassembled and inspected and no problems
were found. Also, the TDCA pump seal water supply from its
discharge was inspected and found clean.
! The swapover of the train A pump to RN system indicates
that at least two of the train A pressure switches
actuated. No swapover occurred on train B which indicates
that apparently only one of the train B switches actuated.
This hypothesis is also supperted by calibration checks
done or the pressure switches, which showed that the three
pressure switches with the highest actuation pressures were
two train A's and one train B. A logic check on both
trains was performed, and no problems were identified.
- - .. . . -- - ._
_ - _ _ _ _ _ _ _ _ _ . . - .
.
_4-
- ,
21
The electrical system performed _ as expected. When more
than one of the train A switches sensed pressure below
decreasing setpoint and the timer timed out, swapover,to RN
occurred. The computer alarm listings indicate that after
51 seconds, all switches in both train A and train B had
reset. Because both trains indicated low pressure, it does-
-
- not appear that an electrical failure was responsible for
the swapover. The wiring of the switches was checked to
ensure that there was no crossover between trains.
Calibration checks of the train A and B pressure switches
were performed. Pressure switches are tubed from separate
manifolds. A failure of a single root valve, manifold, or
pressure switch could not prevent actuation of a separate
train. Although the train B- root valve was found to be
partially clogged, that condition would not affect the
operation of train A. Barton 580 series pressure switches,
- the type used for the -CA system swapover logic, have a
deadband of 10%. It is impossible to predict the exact
reset pressure of a given switch within the deadband, and
the reset sequence of a group of switches might not be the
, same as the actuation sequence.
Therefore, it was concluded that all of.the logic for the
swapover of TDCA pump suction was not made up, which was
proper as was confirmed by several means as noted above.
However, whether a low pressure condition actually
occurred and if it did why train B was not actuated was
not yet known.
An investigation of the hydraulic system was performed to
determine if a low pressure condition actually occurred.
Plant personnel verified that the isolation valves from the
hotwell and upper surge tank were open, however, the CA
condensate storage tank isolation valve 2CA6 was closed.
Normally UST level is maintained greater than 80%; however,
during the event the water level in the upper surge tank
was believed to have been approximately 55-65%. This was
not known until a few days after the event. The tank level
per recorder had stuck at the 90% level and the tank makeup
had been isolated that day for maintenance. Under these
conditic,s, when MDCA pumps 2A and 28 started, the steady
state pressure at the instrument taps was calculated to be
12 to 20 psi higher than the low pressure set points.
Explanations for extremely low suction pressure would be
loss of the normal suction sources due to inadvertent
isolation, check valve failure, or depletion of suction
sources. A failure of the common supply suction check
valve CA129 or the upper surge tank supply check valve CA3
resulting in flow blockage, would be the most probable
occurrence. Valve 2CA1, the notwell check valve, might
have had excessive leakage, short circuiting the supply
from the upper surge tank, subsequently limiting flow to
the CA pumps and causing a low pressure condition.
However, effects of rising hotwell level were not observed.
. - _ . --
-_ __ _ _ __ - - _ __ _ _
,
.-
,.
. .
22
After the incident occurred, and the check valve failure
investigation begun, the common suction supply check valve
CA129 and the upper surge - tank check valve CA3 were
inspected; the discs were observed to swing freely with no
apparent obstructions-or misalignment. The hot'well supply
check valve 2CA1 was inspected : and had virtually -no
leakage. 2CA3 was also inspected and no problems were
observed.
At this point the licensee could not explain why or if an
actual low pressure condition occurred. A dynamic test was
developed to prove operability of the system and obtain
more data to explain the event.
4. Dynamic Testing
In order to obtain transient data on CA pump suction
pressure during CA pump starts, the CA suction header was
'
instrumented at 5 different points with temporary 0-50 t
O.125 psig pressure transmitters and visicorders- .
TT/2/A/9200/14, CA System Autostart Transient Test, was
written and approved to measure -and record suction header
pressure during various CA pump starts. An attempt would
be made to recreate conditions necessary for swapover to RN
. supply on an extended low suction pressure greater than the
3-5 second time delay. In this test, the pumps would be
manually started. The logic for manual starts of the pumps
requires that if an actual low suction pressure existed
longer than the time delay, the pumps would automatically
trip and swapover to RN would not occur. (Swapover to RN
occurs with an autostart signal and extended low suction
pressure.)
The pressure switches were recalibrated to be within their
required settings prior to dynamic testing.
The following conditions were considered in determining
test prerequisites:
a. All 3 CA pumps starting simultaneously would create
the lowest possible transient suction pressure.
b. A 34 second time delay in starting the TOCA pump would
simulate starting conditions under certain actual
demands.
c. Upper Surge Tank (UST) level would be maintained at
60-80% level, as level was estima'ed to have been at
65% during the March 9 event.
d. Starting the pumps with 2CA 6 open and again with 2CA6
shut would show system response with and without the
additional suction head of the CA CST available.
-
~
_
..
4
n 33 .
"
.
23
The test therefore would include 4 manual starts in the
,
following manner:
'
.
(1) Simultaneous start of all 3 CA_ pumps with 2CA6 open.
(2) Simultaneous start of all 3 CA pumps with 2CA6 shut.
(3) Simultaneous start of both MDCA pumps, TDCA pump start
34 seconds later, with 2CA6 open.
(4) Simultaneous start of both MDCA pumps, TDCA pump start
34 seconds later, with 2CA6 shut.
The test was run on March 17 with the unit in Mode 3 and
the following results were observed:
The transients exhibited a characteristic suction pressure
dip of 25-30 psig just after pump start as expected,
lasting 3-4 seconds. (Refer to graph 1.) In each case
where 2CA6 was open, the CA suction pressure never dipped
below the low suction pressure setpoints. When 2CA6 was
shut,- available NPSH was less and pressure was observed to
drop below the low pressure setpoints. This caused control
room annunciators to momentarilyLalarm, but not longer than
the 3-5 second time delay and the CA pump trip did not
occur. In all cases steady state suction pressure was
acceptable.
Af ter the 4 starts with UST level at about 60%, and 2CA6
shut, vacuum on the main feed pump (CF) condenser was
inadvertently lost and the running CF pump tripped. This
caused a CA autostart. Both MDCA pumps started, a swapover
to RN occurred on train A and raw water was again injected
into steam generators A and B. During this event the MDCA
pumps were thought to have started in a staggered manner
1/2 - 1 second apart, which was suspected to have extended
'
the pump start transient past the time delay on train A and
caused swapover. UST level also appeared to be a more
significant parameter than originally believed. Since the
TDCA pump did not start during this event, it was concluded
that it played no role in the swapover scenario.
Evaluation of the results of the initial dynamic testing
and the inadvertant CA start led to the conclusion that the
March 9 low pressure sce..ario was real and therefore could
be duplicated. A further test program was defined and
additional instrumentation installed.
. .
-
..
'
i
A
. .
24
Three (3) starts of the MDCA pumps, staggered by 2 seconds,
were then conducted with 2CA6 shut. UST level varied.from
65-90%. No CA pump trips were observed. On graph 2 one
notes that the transient is indeed extended to about 6
seconds; however, the pressure did not dip below the
.setpoints. The licensee suspected this stagger time to be
too long to cause swapover'and did 3 more starts with pump.
starts staggered at 1/2 - 1 second. In each of these last
3 starts the train A MDCA pump tripped, indicative of an
extended' low pressure. sensed.on train A. The transient is
shown on graph 3. The sequence of starting pump A or B
first did not matter.
UST level was dropped to 55% and the- running CF pump was
tripped to initiate an autostart of the CA pumps. A
swapover to RN occurred ' on train A only. The time
difference between pump starts was measured and concluded
.to be negligible.
1
5. Dynamic Testing Conclusions
As a result of the testing the following conclusions were
reached:
a. With 2CA6 shut and UST level between 55-60%, starting
the MDCA pumps simultaneously or slightly staggered
would cause a swapover to RN on "A" train only because
of 3 reasons.
(1) The suction pressure transient dipped lower and
lasted longer than expected.
(2) The pressure transient as sensed by "A" train
instruments typically lasted 0.3 to 0.4 seconds
longer than "B" train due to the location of the
instruments and the length of impulse lines.
(3) The time delay relay on "A" train (4.15 seconds)
was shorter than "B" train (4.70 seconds).
b. With UST above 90% and 2CA6 shut or with 2CA6 open,
enough NPSH was available during tha trarsient to
pre,ent swapover.
6. Conclusions
Partial swapover of CA to RN on train A occurred because of
the following reasons:
1
1. _ .-...
'
,
'
.; , <
.I
, .,
-25
1. 2CA6 was shut by operations- to eliminate water-
migration between Uni.ts 1 and 2 and to prevent loss tf
CACST level.
'
2. UST : level had dropped to 55-65% level 'without'
knowledge of the operators since the pen recorder _had
stuck at 90%.
3. Two pressure. switches of . Train A suction and one on -
Train B were slightly high out of tolerance.
4. The time delay : relay on Train A was shorter than on
Train B and the transient was sensed by Train A for a
slightly longer time; therefore, swapover only
occurred on Train A.
5. The logic for swapover was made up for the MDCA pump
and not for the TDCA pump.
7. Corrective Actions for Short Term
The licensee has committed to operating with valve ICA6
and 2CA6 open until evaluations determine an operating
envelope for proper autostarts of the CA pumps.
B. Investigation of Clam Fouling of RN Supply
1. History of Licensee Program For RN Fouling
a. On September 3, 1980, Arkansas Nuclear One experienced
low service water flow through the containment cooling
' units due to Asiatic clams in the system. As a
result, NRC issued IE Bulletin 81-03 on April 10,
1981. The Bulletin requested, in part, that holders
of construction permits (Catawba at time of Bulletin)
determine if Asiatic clams were present in the
vicinity of the station, determine whether
infestations were present in potentially affected
systems, components, or systems affected, and outline
corrective and preventative actions. The licensee
responded in July 1981 and March and September 1983.
The following summarizes the responses:
(1) 1981 Response Summary
- The Asiatic clam has been present in the
Duke service area since the mid 1960s. In
1978 Duke formed an ad hoc committee to deal
with clam related problems at all Duke
generating facilities.
,
4 y~. ~_ , , , - , ,, ,, . - - .,
_ _- -
__ _ _ _ - _ - . _ _ _ - _ _ _ _ _ _
,
,
_
. - ..
' '
...
..
.
.
s' ' .
26
,
'
2
-
At Catawba, clams are a-potential ' problem in
4~ two systems that have safety-related ,
implications, the Nuclear Service Water (RN)
system and the Fire. Protection Systems (RF
and RY).
-
The. Catawba RN system includes provisions to
~
prevent the introduction of clams into the
'
system .from the lake via . the RN intake
structure by filtering the water discharged
from each RN pump through a strainer with
1/32 inch openings. Provisions have also
been made to allow backflushing the.
redundant heat exchanger trains and piping
to remove any clams in safety-related RN
'
components and piping. -
-
Chlorinated filtered water is used to
provide normal makeup and maintain pressure
in the fire protection system. The main'
fire pumps on the intake structure are
equf pped with basket strainers on the pump
suction which prevent mature clams from
being - pumped into the system. Periodic
operational testing of the main fire pumps '
will detect any blockage of the pump. suction
veri fy acceptable pump
.
screens and
performance.
-
Periodic performance monitoring programs to
verify adequate flow, and visual inspection
of the intake piping and inlet heat
exchanger heads during maintenance will
provide early detection of any clam .
!
Infestation of raw water systems.
(2) 1983 Response Sr nary
-
Clams are a potential problem in the Nuclear
Service Water (RN) and fire protection (RF
and RY) systems.
-
Following turnover of the RN system to the
Nuclear Production Department, the system
will be functionally tested to verify
capability to supply required cooling flows
in accordance with plant safety analysis.
Subsequently two heat exchangers in the RN
system will be monitored on a quarterly
.
av, ,
y -.-+,gn.. .
w... ..er- 4,,,-7. ,,, n.4y,,-.y.-.p.,,,,.y-. ,.m.,-,g.,,e~, .,,,-..,,,n_,, ,,g.,., ,-a,,v., , , , y -
'.
6
' '
27
basis by setting reproducible flow through
the heat exchanger and recording the inlet
and outlet pressures. The differential
pressure will be checked to determine if
significant fouling has occurred. The heat
exchangers will be a-component cooling water
heat exchanger and the Diesel Generator
Starting Air After Cooler.
-
Preventative actions now being taken consist
of preventive maintenance inspection' of
system components and backflushing of raw
water system piping. The frequency of these
inspections and backflushes is now
determined by what is discovered when
components are examined. System performance
monitoring will also be used to determine
frequency of preventative maintenance once
the plant is operational.
-
The October 17, 1983, final report on the
Corbicula infestation recommends that
flushing of RN lines, especially low flow
lines, be accomplished twice a year.
b. Presently, Catawba has in place several programs and
practices designed to verify adequate Nuclear Service
Water (RN) flow to various systems and components.
This includes the following activities:
-
The RN system is periodically flow balanced to
verify the required RN flow rates are maintained
to the appropriate equipment.
-
Essential heat exchangers serviced by the RN
system are systematically tested to verify their
heat transfer capability, or are cleaned
periodically, based on differential pressure
indications, to prevent fouling. During
cleaning, these heat exchangers and/or any other
raw water system piping are examined for the
presence of clams or unusual fouling conditions.
-
The inspections of the RN system dead leg piping
for clams has consisted of spot radiographic (RT)
inspection. In April 1987, questions concerning
accumulation of clams in the RN system dead legs
resulted in spot RT inspection of three low spots
in each unit's RN to CA lines. In addition, in
March 1988, prior to the Unit 2 CA swapover to
the RN system, two low spots in the Unit 1 RN to
CA dead leg piping were RT inspected. None of
the RT inspections revealed any clams.
.-.
._ . ._- . .. . - . .. --
+
.:
.
l
l
,
- 1
.
'28 i
,
i
2. Flushes and Inspection of Nuclear Service Water (RN) and
Associated Systems
As a . result of the Unit 2 Auxiliary Feedwater (CA) System
Swapover from Condensate (CS) to RN system water and the
subsequent introduction of raw water and clams into the CA
system, the licensee initiated a program of flushes and/or
inspection of dead legs between the RN system and various
. safety-related systems. Since RN water was introduced.into
Train "A" of the Unit 2 CA system, the flushes for Unit 2
included flushing :CS water- through Train "A" CA flow
control valves to steam generators "A" and "B" . The
following summarizes the results of the flushes / inspections
and the NRC's inspection activities:
a. Flushes / Inspections
(1) RN to CA
There is approximately 250 feet of six inch RN
pipe connecting each- train of CA piping to the
main RN Header. This 250 feet of RN pipe as well
as the connecting CA piping contains stagnant RN
water. Each train of both Units 1 and 2 was
flushed at approximately 1500 gpm for greater
than 30 minutes until relatively clean. The
flush path was from the RN header through the
stagnant RN and CA piping to the Condenser
Circulating Water (RC) system. The results were
as follows:
-
Unit 1 - Visual of flush samples taken at
drain valve ICA176 revealed some
small clams (< 1/2 inch diameter)
in train "B" and small clam
fragments in both trains "A" and
"B"
-
Unit 2 - Procedure and sampling identical
to Unit 1: results similar to
Unit 1
After flushing, Unit 2 Train "A" and "B" stagnant
RN piping was spot RT inspected for the presence
of clams. Each train received five spot RTs (17
inches for each spot). The five areas included
the two low spots in each train. No clams were
found.
<
-+
, , r .,
.~ - . . - - , . - . . - . - - - - --
__
. - _ . __ _ _ _- __ _ . __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _
..
.,
.. _.
In . addition to the flushing and RT inspections,
check valve CA172. was disassembled on both
Units 1 and 2 and the CA piping on either side of
the valve inspected for' clams with<a boroscope
~("Videoprobe"). Results were as follows:
-
Unit 1 - Piping for about three feet on
either side of valve ICA172 was-
inspected. Water contained -some
silt and was fairly cloudy. No
clams were detected.
-
Unit 2 - On the downstream side of valve
2CA172, piping was inspected to
valves 2CA116A and 2CA15A. On the
t upstream side of the valve, piping
was inspected to valve 2RN 250A.
'
The water was fairly clear and the
pipe walls appeared clean. No
clams were detected.
Historically, the only flush these lines have
received has been during the quarterly Inservice
Test (IST) of check valves CA171 and CA172.
These check valves - are partially stroked each
quarter in accordance with PT/1(2)/A/4200/08F, CA
Pump Suction Check Valve Partial Stroke Test.
This test requires that isolation valves RN250A
and RN310B be opened and a small flow be
established through the check valves and out
CA183 and CA184 test drain valves to a sump.
Therefore, only minor flushing occurs and RN
water is introduced to the CA system dead legs,
t
During evaluation of the event the following
Unit 2 valves were disassembled for inspection of
the valves and/or inspection of the piping:
2CA1
2CA3
2CA129
2CA172
2CA116A
2CA175
2CA173
2CA174
2CA48
2CA52
2 CAPS 5230
2CA56
2CA60
. ,_ . . _ _ , ,, _ _ _ _ , _ _ ,
'
?; .
- ~ ,
\
. . .
30
No significant problems were identified except
.for valves 2CA56 and 2CA60 which are described
below.
,
(2) Condensate to CA
After flushing RN to CA piping, .the dead legs' in
the CA piping were flushed with condensate ' water
to remove residual RN water from CA lines (both
Units 1 and 2). The flow paths were as follows:
-
CA129 - calla - CA12 - CA15A -Call 6A - CA175
to RC
-
CA129 - CA98 - CA98 - CA18B - CA858 - CA175 ,
to RC
-
CA129 - CA171 - CA858 - CA175 to RC
These flushes were accomplished before the
boroscope inspection described in paragraph (1)
above.
(3) RN to Component Cooling (KC)
RN supplies emergency makeup water to the KC
system. The RN piping between the supply header
and the RN/KC isolation valve is a short section
containing stagnant RN water. The licensee
concluded that clam infestation was likely in
these pipes. RT inspection was performed at
suspect locations (low spots) on both trains of
both units. In addition, A and B trains on
Unit 2 and B train on Unit 1 were flushed. The
flush path was from the RN system through
temporary drains on disassembled valves KC 498
and KC495 to the basement sump. A minimum
pressure of 50 psig was considered adequate to
remove any debris.
The results of the flushes and RT inspections
were as follows:
-
Unit 1 - Train A: RT inspected, no clams,
not flushed.
- Train B: RT inspected before and '
after flush. No clams
before or af ter flush.
Good flow during
flushing.
_ _ _ _ _ -_ - - _ - . . _ - . ._ _ __ ___ __ . _ _ -
_ --___
.
. $. '
s
. .
31
s
-
Unit 2 - Train A: Vertical connection to
RN. -RT before flush
revealed several small
' clams. During flush,
debris could be heard
moving in-pipe - assumed
'
to be clams. RT after
.
flush revealed'no clams.
-1 Flow was good.
- Train B: Horizontal connection to
RN. RT before flush
revealed an object that
appears to be a clam.
The object is still
visible by RT after
flushing. Flow was
good.
(4) RN to Containment Penetration Valve Injection
Water (NW)
RN is the safety related source of water for the
NW system. These stagnant RN lines were flushed
through their normally closed isolation valves to
confirm whether clam . infestation could degrade
their ability to operate. Flow rates were
acceptable with a clean water discharge. Also,
all in-line valves in trains 18 and 2B (NW61B and
RN493) were successfully stroked and checked for
shutoff capability.
(5) RN to Spent Fuel Pool Cooling (KF)
RN supplies emergency makeup water to the KF
system. RN lines to the spent fuel pool are not
restricted by valves with limited flow or
components with narrow passages. Each train has
an unobstructed path to the spent fuel pool.
Therefore, flushing of the stagnant lines was not
considered a high priority or re-start issue.
However, the licensee planned to flush these
stagnant lines shortly after re-start of the
plants. Radiographs taken later did reveal the
presence of some clams and a flush procedure is
being developed.
.
f
l
__ ,- . _ _ _ - . _ . , -.
_ ._- _ _ - -
.-
..
A ^
,
. .
32
'
(6) Condensate through CA to Steam Generator (S/G)
(Unit 2, Generator A and B)
During the initial. event (swapover to RN), flow
through CA train A degraded to approximately 100
gpm to S/G A and 200 gpm to S/G B versus a design
flow' rate of- approximately 300 gpm -per S/G.
Af ter plant shutdown, flow control valves 2CA 56
and 2CA60 were disassembled and found to be-
clogged with clam fragments. Approximately 1 1/2
pints of clam fragments were found in each valve.
The valves are Design ET Fisher Controls valves
with anti-cavitation cages which have 1/8 inch to
1/4 inch diameter flow holes. The clam fragments
easily clogged the small holes. TDCA Pump Flow
Control Valves (2CA48 and 2CA52) were
disassembled and inspected. No clams were found.
Later investigacions revealed that only A Train
motor driven CA pump swapped over to RN.
Due to RN water and clams being introduced into A
Train MDCA pump suction and discharge piping, it
was necessary to assure that these lines were
free of clams and clam fragments that would l
degrade flow. The lines had. seen. approximately
20 minutes of degraded flow from condensate after
switch back from RN to condensate. Af ter the
flow control valves (2CA 56 and 2CA 60) were
disassembled and cleaned, the A train MDCA pump
suction and discharge were flushed approximately
10 minutes to each S/G using condensate water.
The flow path was from condensate through the CA
pump to the S/Gs. The flow rate was approxi-
mately 700 gpm, or about twice the design flow
rate. Therefore, the suction piping saw 700 gpm
for approximately 20 minutes. There was no
degradation of flow. The valves (2CA 56 and 2CA
60) were disassembled and inspected for clams a
second time. A few small fragments were found in
each valve. Further investigation indicated
that, based on the location of the small
fragments and the method used to clean the
valves, the fragments most likely had not been
removed during the first cleaning. The licensee
declared the lines clean and reassembled the
valves.
In addition to the above flushes, piping down-
stream of CA pump 2A was inspected for clams at
valves 2CA 27, 2CA 28, and 2CA 29. No clams were
found.
>
, =- ---- - . , . , . , - - - , , , - - , - , , - - - - - --
%
- s-
.- ,
y
Additional flushing . occurred during flow
balancing the CA system. Train "A" was operated
a minimum of 20 minutes at approximately 600 gpm
through the pump- and 300 gpm to each S/G. No
flow degradation occurred.
After all flushes had been completed, the CA
system again swapped over to the RN system during
dynamic testing (see paragraph III.A.4) in mode
3. After switching back to Condensate Water,
flow control valves - 2CA56 and 2CA60 reacted
properly and therefore no further inspection was
required.
b. NRC Inspection Activities
The NRC AIT members reviewed / observed the following
relative to the activities detailed in paragraph a.
above:
(1) The team reviewed flush paths, completed
procedures and results of the flushes described
above. The procedure review specifically
included the following procedures:
TT/1/A/9200/13 - RN to KC Assured Source Piping
Flush
TT/2/A/9200/12 - RN to CA Piping Flush
TT/2/A/9200/18 - RN to KC Assured Source Piping
Flush
TT/2/A/9200/12 - RN to CA Piping Flush
The RN to NW flushes were documented on a
memorandum to file dated March 12, 1988. The
team questioned whether a memorandum was adequate
documentation for this work. The licensee
pointed out that no plant conditions were changed
and that valve manipulations for the flushes were
documented under their Removal and Restoration
(R&R) procedures. The team verified by visual
observations that systems were restored to their
correct alignment after flushing.
(2) For the RN to KC flushes, valves KC495 and KC498
were disassembled. The valve work was covered by
Work Requests 5289MNT, 5290MNT, 5291MNT and
5292MNT. The team reviewed the completed Work
. _ -
_ . , , ,_ _ _ _ .
._.
5 S
.
2..
' 'e .
Requests including associated ' documentation to
insure that valves. and systems were returned to
-
required conditions.
(3) Video tape recordings of the boroscope inspec-
tions performed of CA piping through valve CA172
'were reviewed.
(4) The two trains of Unit 2 stagnant piping (250
feet / train) between the RN header and the CA
system were walked-down to insure that low spots
were identified for RT inspection. RT film for
these sections of pipe were reviewed.
(5) RT film associated with - the KC flushes were
reviewed.
(6) Valve lineup and preparations for Train.1B RN to
KC flush was observed.
(7) Operability evaluations (PIRs 1-C88-0108 and
2-C88-0107) relative to flushes were reviewed.
3. Evaluation of Short Term Corrective Actions
(a) Effectiveness of Short Term Actions
The licensee's short-term corrective actions to eliminate
clams and clam shells from the CA system lines and the RN
system lines by flushing these lines with relatively high
volumeteric flows while obtaining confirmatory indications
that clams and clam shells are absent from these lines was
judged adequate for the restart of Catawba Units 1 and 2 ;
based on fundamental fluid dynamics considerations and
confirmations by visual and radiographic examinations.
However, long term corrective actions are also required to
.
assure the effective control of clams in the RN system
lines as the present configuration of the RN/CA system
interfaces are conducive for clams growth and infestation.
-
A brief discussion of these considerations is given below.
(b) Flushing of CA System Lines
The CA system lines were flushed with a flow rate of
approximately 700 gpm. This corresponds to an average flow
velocity of 17 ft/see to 31 ft/sec in the 4-inch and 3-inch
diameter lines, respectively. The CA pump discharge line
rises from the CA pump elevation to the steam generator
,
d
- . . - - - , , .,. ,v.- , -- -
--. ---- - - . . - . . - - --
.. .
- a
>
. .-
35
level. The Reynolds numbers associated with these flow
velocities are of the order of 4x10' to 8x105, well into
the turbulent flow regime. Empirical velocity profiles
indicate that, within a distance of a few hundredths of an
inch from the interior wall of the pipe, the flow velocity
is already close to 60% of the average velocity. Based on
empirical data and drag and lift coefficients, this range
of flow velocities is estimated to create sufficient drag
and lift forces on the clams and clam shells in the CA
system discherge lines to flush them out of the lines.
As to the 10-inch diameter suction line to the CA pump, it
runs vertically down from the RN line. The average
velocity in the line is of the order of 3 ft/sec for a flow
rate of 700 gpm. The flow is still sufficiently turbulent
so that the drag and lif t forces on the clams and clam
shells are estimated to be sufficient to flush them out of
the line.
The .above discussions indicated that the short-term
corrective action of flushing the CA lines with a 700 gpm
flow is judged adequate for the removal of clams and clam
shells from the lines. This judgement, is confirmed by the
observatior. that the number of clam shells and shell debris
collected at the CA flow control valve was practically none
after flushing.
(c) Flushing of RN lines
The RN lines were repeatedly flushed with a flow rate of
1500 gpm. This corresponds to an average velocity of
approximately 17 f t/sec in the 6-inch diameter RN lines.
The Reynolds number associated with this velocity is about
8x105, well into the turbulent flow regime. As indicated
earlier, for turbulent flow inside a pipe, at this Reynolds
number, the velocity at a distance of a few hundredths of
an inch from the interior wall of the pipe is already close
- to 60% of the average velocity. Therefore, the turbulent
! flow is estimated to create sufficient drag and lift forces
- on the clams and clam shells to flush them out of the NSW
i lines. This co,'clusion is supported by two confirmatory
indications: (1) the results of visual inspections of
selected portions of the RN lines indicating the absence of
I clams and clam shells, and (2) the results of rad'ographic
examinations of selected portions of the RN lines including
all the low points indicating the absence of clams and clam
l shells.
l
l
I
.
L
.w
a
. .
36
4. Long Term Corrective Actions
The presence of Corbicula sp. organism or shells in the local
environment of Catawba Station (Lake Wylie) was well recognized
by the licensee. In response to NRC IE Bulletin 81-03, "Flow
Blockage of Cooling Water to Safety System Components by
Corbicula SP. (Asiatic clam) and Mytilus SP. (Mussell)," actions
were taken by the licensee to address clam-related problems.
These actions included provisions in the FSAR section 9.2.1.6 to
prevent the introduction of clams into the RN system from the
lake by filtering the water discharged from each RN pump through
a strainer with 1/32 inch openings; provision to allow
backflushing the redundant heat exchanger trains and piping to
remove clams from safety-related components; provisions of flow
elements in the RN system to allow verification of adequate
service water flow to safety-related heat exchangers during
performance monitoring programs; and the visual inspection of
the intake piping and inlet heat exchanger heads during
maintenance for the early detection of clams.
These actions collectively have been effective in controlling
clam-related problems associated with the component cooling
water heat exchangers and diesel generator cooling. However, :
due to the relatively stagnant conditions in the 250 feet of
6-inch diameter segments of the RN to CA system lines,
additional corrective actions are required to assure that clams
will not grow in these lines.
The 250 feet of 6-inch diameter piping from the RN system to the
CA system is not entirely stagnant, but has a flow of about 10
gpm, and is likely to have temperatures above 65 F. These
conditions provide a favorable environment for the breeding and
growth of the C_oribicula sp. There are several options for the
control of the breeding and growth of the species:
Option 1: Chlorination
Shock chlorination to control mature clams requires a high
concentration of free residual chlorine of 10 to 40 ppm for a
duration of about 54 hours6.25e-4 days <br />0.015 hours <br />8.928571e-5 weeks <br />2.0547e-5 months <br /> to ensure a 90% mortality. To
control shelled larvae of about 200 microns in size, chlorine
residual of 0.3 to 0.4 ppm is required for a duration of about
100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> to ensure a 100% mortality. Since clams grow about
1/4 to 1/2 inch per year depending on the environmental
conditions, to control the size of clams in the system to less
than 1/8 inch would require at least a quarterly treatment. The
effectiveness of the treatment can be enhanced by performing it '
during peak clam spawning periods, and the use of clam traps and
mechanical cleaning, if feasible, during outages. Chlorination
is generally acceptable only in closed systems.
.. _ -
t, ..
4
. .
37
Option 2: Heat Treatment-
Corbicula sp. is more vulnerable ' to heat than to chlorine
treatment. A 2-minute exposure to 120 water would lead to a
99% mortality of mature clams and larvae. Another nuclear power
plant licensee has observed a 100% mortality rate when their
service water system was flushed ~ with 170 F water from the
auxiliary boiler for about 30 minutes. The backflushing of a
system with sufficiently hot water will achieve the dual
objectives of ensuring a 100% mortality of mature clams as well
as larvae, and the removal of clam shells from the system,
Option 3: System Modifications
Potential system modifications may involve the removal of the 10
gpm flow through the relatively stagnant RN line by re-routing
the 2 inch line to another location; relocating the CA/RN
interface valve or adding another isolation valve upstream of
the relatively stagnant RN line; or provide means for
backflushing the relatively stagnant RN line. These system
modifications aim at developing an un-inhabitable condition for
clams or allowing the line to be chlorinated, heat treated, or
backflushed. t
Option 4: Other biocides or Chemical Treatments
On a longer. term, other biocides or chemical treatments may be
developed to inhibit and control clam growth as well as meeting
existing state environmental restrictions. Currently there are
many questions about the feasibility of developing an effective
program, and its testing and verification. Therefore this
I
option appears to be suitable only for a much longer term
application.
IV. REGULATORY REQUIREMENTS
A. Catawba Technical Specification 3.7.1.2 requires that three
independent steam generator auxiliary feedwater pumps and .
'
associated flow paths be operable with:
-
Two motor-driven auxiliary feedwater pumps, each capable of
being powered from separate emergency busses, and
-
One steam turbine-driven auxiliary feedwater pump capable
of being powered from an operable steam supply system when
the unit is in modes 1, 2, or 3.
With one auxiliary feedwater pump inoperable, it is required to
be restored to operable status within 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> or be in at least
hot standby within the next 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> and in hot shutdown within
the following 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />.
'
_ .
..
g.
' '
38
With two auxiliary feedwater pumps inoperable the unit must be
in at least hot standby within 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> and in hot shutdown
within the following 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />.
It is further required that each auxiliary feedwater pump be
demonstrated operable at least once per 31 days on a staggered
test basis by:
-
Verifying that each motor-driven pump develops a total
dynamic head of greater than or equal to 3470 feet at a
flow of greater than or equal to 400 gpm,
-
Verifying that the steam turbine-driven pump develops a
total dynamic head of greater than or equal to 3550 feet at
a flow of greater than or equal to 400 gpm when the
secondary steam supply pressure is greater than 600 psig
and the auxiliary feedwater pump turbine is operating at
less than or equal to 3800 rpm and,
-
Verifying that each non-automatic valve in the flow path
that is not locked, sealed, or otherwise secured in ,
position is in its correct position.
It is further required that at least once per 18 months that the
valve in the suction line of each auxiliary feedwater pump from
the Nuclear Service Water System be tested to verify that it
automatically actuates to its full open position within less
than or equal to 15 seconds on a loss-of-suction test signal.
The conditions of the plant which may have been contrary to the
above requirements are:
1. On March 9, 1988, with the unit operating at 20% power,
three independent steam generator auxiliary feedwater pumps
and associated flow paths were not operable in that;
-
During an actual turbine trip /feedwater isolation
transient and resulting auxiliary feedwater (CA)
initiation, the train A discharge flow control valves,
,
2CA 56 and 2CA 60, became clogged with Asiatic clam
shells and did not properly function af ter MDCA pump
2A suction realigned to the nuclear service water
assured source.
B. 10 CFR Appendix B, Criterion XI, requires that a test program be
established to assure that all testing required to demonstrate ,
that structures, systems, and components will perform
'
satisfactorily in service is identified and performed in
accordance with written tests procedures which incorporate the
requirements and acceptance limits contained in applicable
design documents. The test program shall include, as
i
- -- -
. ,_ , . . . _ . - . _ _ . - . _ _ _ , _ _ _ . , ,_ _ __.-._.. -.___. . - . - -
.
. .
.*
- '
39
appropriate, proof tests prior to installation, pre-operational ,
tests, and operational tests during nuclear power plant
' operation, of structures, systems, and components. Test results
shall be documented and evaluated to assure that test
requirements have been satisfied.
The surveillance program recommended in response to NRC
Bulletin 81-03 stated that flushing of RN lines, especially low
flow lines, twice a year was appropriate. This was documented
in a memo dated October 17, 1983, from J. J. Hall to
R. W. Quellette and was presented to the AIT.as the final report
referred to in the early 1983 Bulletin responses that was to be
available by November 1983. In lieu of flushing, the licensee
performed RT of selected pipes for clam infestation.
This flushing was not being accomplished. Also, the
radiographic inspections that were made were inadequate in that
they did not show clams and clams were present.
V. CONCLUSIONS
A. Reportability Review
The inspector reviewed the telephone reporting of the event to
the NRC. The licensea appropriately reported the Unit 2 Reactor
Trip as required. L;ter that evening, the licensee recognized
the possible significant degradation of the CA system. Although
the 11censee communicated the additional information effectively
to NRC: Regior. II the following morning, it would have been
appropriate to initiate a followup phone call to NRC on the
previous backshift in order to allow a more timely assessment by ,
> the NRC of the event. The licensee indicated that personnel
would be sensitized to make followup calls when bdditional
significant information is discovered relative to events and
that more procedure guidance would be developed in this regard.
The next day, the licensee properly reported the shutdown of
both units and the Unusual Event caused by the degraded systems
during shutdown.
B. Auxiliary Feedwater Suction Swapover to RN
1. CA to RN swapover occurred, during the March 9, 1988 event,
to the A train MDCA pump due to a momentary low pressure
condition caused by the suction lineup being different than
on previous CA pump auto-starts. The CACST was isolated ,
'
and the UST level was lower than normal.
2. A more timely debriefing of the opera + ors involv0d would
4
have expedited the event evaluation,
i
l
v
-s-- -
-- -e ---- , , - - - . ,.y. , , , ,m., --
, - , -..,,,,-,w-. , _ - , , . . . - , , _ - - -,...-.,,._,--------.c.---- , - - - --+=2-
~ . . .. . - . . . . . .
-
, 1. .:
a'
.
". .
40
13. There were 'no precursors to the event which would have
provided ^ early - warning and no' equipment problems were
repetitive or were ignored.
4. Short: term corrective action is adequate.
C. Auxiliary. F' eedwater Degradation of Flow
1. The' CA flow degradation was due to clam particles clogging -
the flow control valves. The clam particles were induced
when the CA suction swapped to RN supply. The clams in the-
RN supply were there due to inadequate surveillance of the
stagnant lines.
2. Some surveillance for clams in the stagnant lines was
conducted by radiography but was inadequate as clams were
found in the lines after radiography.
3. The licensee took appropriate and timely corrective actions
by evaluating similar conditions on the other units.
Catawba Unit I was shutdown and the stagnant lines flushed
and inspected. Stagnant lines at McGuire Units 1 and 2
were inspected by boroscoping.and they were found clean.
i 4. Flushing of stagnant lines at Catawba was recommended but
not accomplished.
'
l 5. -The short term corrective-actions are adequate.
[ VI. EXIT
- .
.
,
The inspection scope and findings were summarized on March 18, 1988,
with those persons indicated in section I.V. The inspector described
the areas inspected and discussed in detail the inspection findings.
No dissenting comments were received from the licensee, but areas of
concern will be addressed at the forthcoming enforcement conference.
[
'
.
The licensee did not identify as proprietary any of the material
i provided to or reviewed by the inspectors during this inspection.
i
l
l
i-
!
h
.
)lIl.
.
.
,
.
, '. .
.
.
a
s
r7. o. d
s .
o on-
' . '
o -
'3 p, T r
s_ e<
c .
-
?. _4
. .
s
p
. _
< _
>
l
S
B
.
f J
o 0 .
1 5., - - h ~
- - - . o
~. -
- - ;
e.
g
a
P
- .
-
f l
l
I
-
.
.
-
l
1
P
I
R
T
P
U
P
O
N
M
^
W i
.
.
.
.
.
,
.
-
T
N 1T -
E U .
t
f HH ,
H PS
C A .
A R6 .
T G. A
T C
A 2 ,
-
.
.
-
S -
P
M
U
P ,
A
C
3
l
O
'
m
N
.
..- .- - , _ - - .
. . - - . - - ..-..- . _. ._ -.. . . .,
'4.*..
,
^
i-
Atiachment 1 GRAPH 2- _ .
Page '2 of .3' ~f.
-2.SECOND STAGGEP,ED START OF f1DCA PU!!PS - 2CA6 SHUT-UST 60% .' ~ '
'
NO PUMP. TRIP '
~-
-
.
n
. L
l
!
l
i
I
, .$ t y s e
--
BTre.h
.
,-
'
'
I --
A Tr.!a 'l
' \( --
--
.. .
-
-. --
1
,.o -.
.N
=
mse
- /
..
I
a.>
< M .,
L
i t -- t - -- - l - -- I - - I -- 1 I - t- -
sesoak
L
y ,,- --,, . , , , -
--,,w....,,
.
. . . , ,,, ,, w v --- n , , -m, , , .- -. -.-e. N.-- n--y . - - . , . , , , , , ~
-
,--,,,+,,------av, , . , , . , - ,, -
, a n v. ,, , , , -,,
, w-. _ . , - -. ,_- .
~-
_
- -
_-9
'
AttadEnent 1 GRAPH 3 ,
Page 3 ofL3; . ..
l 1 SECOND STAGGERED START,OF MDCA PUMPS-2CA6 SHUT-UST 60%-A-MDCA PUMP TRIP ~ d N e-
,
e e-
?
- , -i
_
'T{
. . ,
-
m
0
t
r ~ 5o esg
'~
B %;u .
'-
l -
%'
! . ::
-- --
Sop 3.g
--
l -_
-
m --
-
l ,_
AL a
'
Qy
-. __
.
-o
__ __
eillb
$.
-- O
i t -- 6 ; ;- ;
,--
_,
-sea.-45
'
, _
-
a ,---,,,, ,. -- +,,,., ,, -, -, e ,v ,e.. -~, c ---r---.--g ,- - - . . - , , m -- , r- ,, - ,,.-e-,
- _ _ _ - _ _ _ ______
k
. s
ATTACHMENT 2 Page 1 of 2
0
o
OETAIL FIGURE 1
, . Dgsigns ED, E AD, ET, E AT -
Installation 7. Iron easy.e valve bodies are rated at 125 and 250 lb.
ANSl; steel and alloy steel bodies are rated at 150, 300 and
1. Before ,nstalling
i an easy-e body, inspect it for any sh,p. i
600 lb. ANSI. Do not install the valve in a system where the
ment damage and for any foreign material that may have working pressures exceed those specified in the standards.
collected during crating and shipment. .
2. Blow out all pipelines to terrove pipe scale, welding
Maintenance
slag, and other foreign materials.
3. Install the valve so that flow through the body will be WARNING
I in the direction indicated by the flow direction arrow on
the body. To avo:J personal injury and damage to the
process system, i:of t.te the control valve from
4. Install the valve using accepted piping practices. For the system and release all pressure from the
f flanged bodies, use a suitable gasket between the body and body and actuator before disassembling. '
pipeline flanges. Before installing welding-end bodies having
composition seats, remove the trim to avoid damage to elas-
tomer parts fron: heat generated by welJing. Disassembly
Part names used in the following steps refer to figure 2
5. Control valves with an easy e body can be installed in except where indicated.
any position, but the normal method is with the actuator
vertical above the body. .
1. Af ter the actuator is disconnected and taken off the
.
body, remove the nuts (key 16, figure 7) or cap screws from
6. If cont;nuous operation .is required during maintenance the bonnet flange.
and inspection, install a conventional three valve bypass
around the body. 2. Lift off the bonnet along with ' 'ug and stem
$1140 , ,
1
. Y,
-
PACRING Ft ANGt
y ;E ( '
DON %E7
-
P ACK INQ
BONNETGARET
g q
g.w w
.,
N BACKUP RING
Q Fu
SPIRAL h0L%D GASEE ?
N
-
-
:N
w % F4,_ y 51 AL RING
CAGEGAMET -g b '-
%R g we
1 }?
-f.'y -i,
/ p VALYs PttF1
/
~
~
~s
'
~ CACE
GROOvEPm -' #
r - St&TAWG
'
.*f" D.
o ~ see
m qr
St AT RING GAEsti . -
.
k.
Fmure 2 Sectional View .1 De*ign ET %fve Bc4 with Full Size Trim
2
l
_a
. _ _ - _ _
o' ' Attach'nent 2 DETAIL FIGURE 2 Page 2 of 2
*
,
Designs ED, EAD, ET, EAT
t
.
.
.e
' D 2] 3
'
* EL,
k gt
*
, -
;
*""
iy r ( < .J -
QUICK OPENING EQUAL PE RCENTAGE C AVITROL@ l
.
i
- . 1
[ LINEAR WHISPER TRIM 31
Figure 3. Cages for easy-e Bodies
3. Loosen the pacWng flange nuts (key 5, figure 12) and 6. Lif t out the seat ring and its gasket. For composition
y separate the valve plug and stem from the bonnet. If th) seats,lif t out the disc retainer, disc seat and disc. r
stem needs replacing, drive out the groove pin and unscrew
,
the stem. If the valve plug needs raplacing, always reclace
the entire valve plug and stem assembly. Reassembly
Part names used in the following steps are shown in figure 2
except where indicated.
1. Clean all gasketed surf aces and use all new gaskets for
Never use an old stem with a new valve plug. reassembly.
The use of an old stem <equires drilling a new
groove pin hole through the stem, thereby 2. For restricted trim (figure 8), replace the seat ring
weakening the stem. adaptor gasket (key 14) and adaptor.
4. The internal parts of the bonnet can be disassembled,if 3. Install the seat ring gasket and replace the seat ring. If a
desired. For packing replacement, instructions are given in composition seat is used, assemble it by inserting the disc
the section "Replacing Packing." For bellows seat replace- into the disc retainer and slip this assembly over the disc
,
ment, follow the section entitled "Replacing Bellows Seal." seat.
4. Install the cage on top of the seat ring. Be sure that the
CAUTION
l' cage slips onto the seat ring properly. Any rotational orien- l tation of the cage with repect to the body is acceptable. I
The exposed portion of the cage provides a
guiding surf ace that must not be damaged dur. 5. For full site trim, place cage gasket, spiral wound gas.
ing disassembly or maintenance. If the cage is ket, and bonnet gasket on the shouldcr of the cage,
stuck in the body, use a rubber mallet to strike
( the exposed portion at several points around 6. For restricted trim, place the cage gasket, spiral wound
the circumference. gasket, and another cage gasket on the sholder of the cage.
'
For 2" x 1" angle and 11/2" x 1" globe bodies, a body
3 5. Lif t out the cage and gas,ats. For restricted trim adaptor gasket (key 20, figure 11) is required be* ween the
"
(figure 8), also remove the cage adaptor (key 4) and seat cage adaptor and body. Insert the cage adaptor and place
ring adaptor (key 5). the bonnet gasket on the adaptor.
1
- - - - - - - - - - - - - - )
'
.
* *
'
ATTACHMENT 3 _
Page 1 of 3
UNIT E D ST AT Es
,[pMf 'o
" ,d
NUCLEAR REGULATORY COMMISSION
&' REoloN 11
C' o 101 M ARIETTA sTRE ET. N.W.
j
s
(* ATLAN1 A, GEORGI A 30323
' \ $'
% 3l)3l)Y
Oocket Nos. 50-413, 50-414
License Nos. NPF-35, NPF-52
Duke Power Company
ATTN: Mr. H. B. Tucker, Vice President
Nuclear Production Department
422 South Church Street
Charlotte, NC 28242
Gentlemen: .
SUBJECT: CONFIRMATION OF CONCURRENCE - STARTUP OF UNIT 2 - DOCKET NO 50-414
This letter is to confirm our concurrence with your intention to restart Unit 2
following completion of your short term corrective actions relative to the
Catawba Unit 2 trip on March 9,1988 and subsequent swap-over of the auxiliary
feedwater pumps suction supply and degraded auxiliary feedwater system flow.
We had earlier confirmed with you our understanding of commitments to be
completed prior to restart of Units 1 and 2 in a letter dated March 11, 1988.
Af ter a review of your short term corrective actions by the NRC Augmented
Inspection Team we gave you verbal concurrence to restart Unit 2 on March 18,
1988 at about 9:00 a.m. Concurrence was given with the understanding that the
Auxiliary Feedwater (CA) Condensate Storage Tank (CST) isolation valve (CA-6)
would remain open pending your analysis of CA system transient data. Also, it
is our understanding that if your short term analysis of CA system transient
data supports a change in system configuration to assure a supply of quality
condensate grade water you will keep NRC Region II and the Catawba Resident
Inspectors informed.
If your understanding is dif ferent from the above, please inform this office
promptly.
Sincerely,
t
J. Nelson Grace
Regional Administrator
cc: J. W. Hampton, Station Manager ~
Senior Resident Inspector - McGuire
S S // 4 /1/ p 4 6 / g
p - - - _ _ _ _ .
- -
3
Attachment 3 Page 2 of 3
UNITE 3) STATES
l /ja ano "
NUCLEAR REGULATORY COMMISSION
*
, 3" REGION il
**j 101 MARIETTA STREET,N.W.
* I
f ATLANTA. GEORGI A 30323
%.,,,,*/
t March 11, 1988
Docket Nos. 50-413, 50-414
License Nos. NPF-35. NPF-52
Duke Power Company
ATTN: Mr. H. B. Tucker, Vice President
Nuclear Production Department
422 South Church Street
Charlotte, NC 28242
C:'tlemen:
SUBJECT: CONFIRMATION OF ACTION LETTER - DOCKET NOS. 50-413 AND 50-414
This letter is to confirm our understanding of comitments made during a
telephone call between Mr. T. Owen, Catawba Plant Manager, and Mr. T. Peebles
of my stati on March 10, 1988. The telephone discussions related to the
Catawba Unit 2 reactor trip and subsequent swap-over of the suction to the
auxiliary feedwater pumps and degraded auxiliary feedwater flow. Subsequent
discussions were held between our staffs to fully understand your actions
relative to the Notice of Unusual Event and shutting down of Catawba Unit 1 as
follow-up actions to the Unit 2 degraded auxiliary feedwater flow circum-
stances.
With regard to this situation, it is our understanding that you have comitted
to notify the NRC Region II and Senior Resident Inspector, with adequate
advance notice in order to obtain the concurrence of the NRC Region II Regional
Administrator or his designee prior to startup (Mode 2).
If your understanding of this matter differs 'ren that stated above for either
unit, contact this office promptly.
Sinceirely,
jG J. Nelsen Grace
Regional Administrator
. h
CAL 50-413-8801
50-414-8801
cc: T. B. Owen, Station Manager
Senior Resident Inspector - McGuire
-wp3 f 94WF
lp
- _ - - - - - - - - - - - - - - 1
r ,
.,.
' Attachment 3 Page 3 of 3
.
$7.rfou UNITED STATES
o NUCLEAR REGULATORY COMMISSION
,a ' #
REGION 11
3
**
j 101 MARIETTA STREET. N.W.
ATLANT A, G EORGI A 30323
t .
'+, ,d
.....
MAR 17 ;553
,
Docket Nos. 50-413, 50-414
License Nos. NPF-35, NPF-52
Duke Power Company
ATTN: Mr. H. B. Tucker Vice President
Nuclear Production Department
422 South Church Street
Charlotte, NC 28242
Gentlemen:
SUBJECT: CONFIRMATION OF CONCURRENCE - STARTUP OF UNIT 1 - DOCKET NO. 50-413
This letter is to confirm our concurrence with your intention to restart Unit 1
following completion of your corrective actions relative to declaration of an
Unusual Event and shutting down of Unit 1 on March 10, 1988. We had earlier
confirmed with you our understanding of commitments to be completed prior to
restart of Units 1 and 2 in a letter dated March 11, 1988. After a review of
your corrective actions by the NRC Augmented Inspection Team we gave you verbal
concurrence to restart Unit 1 on March 11, 1958 at about 11:00 p.m. Also, as
stated in our March 11, 1988 letter, you have connitted to obtain NRC Region 11
concurrence prior to Unit 2 restart.
If your understanding is different from the above, please infonn this office
promptly.
Sincerely,
9
[ J. Nelson
Regional Grace
Administrator
j
cc: T. B. Owen, Station Manager
Senior Resident Inspector - McGuire
i l l l t
[(, r
e
*
.. ATTACHMENT 4 Page 1 of 3
'*
, CA FLOW DIAGRAM
-
a
-
. .. . :
? <
m
e 1 ,.
L Lt
_ s s
5 s I
g !.-
_
,,- ,, .,
.
4
Y a
m s -
y ~l"n
I
- -
5?
" I
- ~ t
il
-
aC
e
C a
e i
e3
$k
J,
-c>o-N :
"-
:. C ne
C ;
U *O
.\
s q >< ><
-
i
C f 5
9. .y sn O--
EC ; 9.. n g
G
,
e
-
J sy
9 , l 5
sf .
g
e-
.
1jg -
1, _
se --e e
8
E @; .
S O- -
X 4~e
?
d
Et
VV
o-m o-v e-m
-
i
,._ ..
,
g s
1 = 21 1
r, : s a
-
: : s a
30-- C s *C * :C -
* 'C *
-
s s 3 : a 3
e s
3 a
3
'
aC a ,,a s0 nC s * C na 70 n C
C X 3 0--X
i C X C X
i*
9-4 9-@ 9-4 9-R
__ __ __ __ __ __
__ __
( s. "_) (s. 3 (s D ( s. D
__.
w
.
. .
*
f
, cu
o
[
ro
5-
V
:
CA I
CO @ STORAGE TK
42.'les gal
38 OPH OVERFLOW
u r n
>
n
m
ICS73 C(M)ENSATE -4
i L N ETURN 0
TO CST 1 ,
UNIT SUPPLY
I'
_
ICS63
>
X ICS68 X Q>
ICS74 3l:
ICS74 COWENSATE
1C567 C : RETLADe
UNIT 2 _ TO CST 2 SUPPLv -
00@ENSATE COWENSATE
SUPPLY 7C SMY
FROM UNIT 3 FR(M tmli 2
I
'a'
ro
o
-%
u
i
TITLEe NOTESs
10 W' CN-CF-CA-3 l DATE' 3-7-86
AUXILI ARY FEEDWATER ( CA) CACST SUPPLY / OVERFLOW NORMAL REF.
LINEUP D"" 8 LY / CM l " * *'
TRAINING USE ONLY
_
g- .
S.,,l ;l $nA $ a R m g n m g g E a
! ,$ w "
I*".#
$
Y
'
'
d' 6
8
V A'l -
g,
b
N( 7
l -
3 Y_
fA 'Y X y 4
f
RL g O
C-
t O e-
'D N' , E
S
oT
@f
l
.
L A
A
p@E
<
L
2
g
U
G
- -
-
M - - A N
I
Rf -
o C N
o
l
.
- I
A T - F j A
S
U C y R
6 N
C
-
,
T
y
-
G 0'
N F
.
g
-
2
4
4 m 0
8
E
R m
6
L
a
E
F
O
S
E .
C
R
6 U
. O
. S
S
.
,
'
EP
D D
At
._ _.
T
S
C ] RP
G
A
A EC
_ C .
T
_ AO
ST
N
_. ER - .DE
r SNT
E
3 TOA
B OCW
N
L i
E
)
A
C
(
R
E
* T
4 A
Y W
6 D
3 E
. E
L F
E 2
Y
R
A
I
. .
m=w-
e
v
i
.
u
w
e I
EX
TU
L
IA
T
L
ea
r
e.
_
I
_ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ - }}