ML20235K820
| ML20235K820 | |
| Person / Time | |
|---|---|
| Issue date: | 01/31/1989 |
| From: | Padovan L NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | |
| Shared Package | |
| ML20235K761 | List: |
| References | |
| NUDOCS 8902270256 | |
| Download: ML20235K820 (10) | |
Text
_ _ _ _ _
f,}i
~f
- r"
.7
-; N.. ;
a ',.. :
AE0D/T902.1 TECl!NICAL REVIEW REPORT INADVERTENT REACTOR TRIPS DUE TO RCS FL0ll INSTRUMENTATION MAINTENANCE ACTIVITIES..
January 1989 i-
' Prepared by:
L. Mark Padovan Office for Analysis and Evaluation of Operational Data f-U.S. Nuclear Regulatory Commission I
8902270256 890207
{DR TOPRP EMVWEST PDC l
SUMMARY
A review of recent plant operating events indicates reactor trips have been inadvertently generated at several Westinghouse pressurized water reactor plants while at power, as a result of maintenance activities on reactor coolant system (RCS) flow transmitters. These reactor trips occurred at Trojan on August 16, 1988 (Ref. 1), Sequoyah Unit 2 on May 23, 1988 (Ref. 2),
Salem Unit 2 on April 21,1988 (Ref.3), and Diablo Canyon Unit 1 on January 8, 1988 (Ref.4).
In each case, the reactor trips were attributed to personnel error.
Technicians failed to follow established procedures, and generally did not understand that with the plant operating at higher power levels, even careful manipulation of RCS flow transmitter isolation, calibration and vent valves could create sufficient perturbations in sensed RCS flow to cause inadvertent reactor trips.
Flow transmitter sensitivity to maintenance activities is a result of system design.
Three flow transmitters in each RCS lcop share a ccamon "high side" pressure tap. During maintenance activities on a single transmitter, perturbations induced at that transmitter can be propagated to all three transmitters, satisfying coincidence to generate a reactor trip signal.
Unplanned reactor trips cause unnecessary challenges to plant safety systems, and provide undesired opportunities for unexpected equipment malfunctions and operator errors. Accordingly, both the NRC and the nuclear industry have focused increased attention on the issue of unplanned reactor trips at U.S.
plants. Nuclear Steam Supply System (NSSS) vendor owners groups have also established programs to address reactor trip reductions.
DISCUSSION 1.
Backgrouna Information Measurement of RCS icop flow in pressurized water reactors is typically accomplished utilizing differential pressure transmitters. These transmitters are installed across an elbow in each reactor coolant loop and measure pressure drcp developed by RCS flpw through the elbow (see Figure 1). At Westinghouse reactors, three transmitters in each loop share a common "high side" pressure tap and tubing run from the RCS. The low pressure side of each differential pressure transmitter is connected to the RCS by separate tubing. Typica lly, an isolation valve is provided in each of the low pressure side sensing lines, but only one isolation valve is provided in the common high side sensing line.
At plants with this design, the high side line cannot be isolated when the plant is above the P-8 setpoint (generally set at 35 to 39 percent reactor power for Westinghouse plants), since the high side line is cenmon to all three flow transmitters and isolation will cause a reactor trip. Accordingly, maintenance activities are rormally performed on the transmitters without isolating the high side sensing line. On some plants, however, isolation valves are provided on the high side leg of each flow transmitter to facilitate maintenance.
v.
(
A reactor trip will' occur when a low flow signal is generated on two out of three channels, if two out of four nuclear power range channels exceed the reactor protection system P-8 setpoint.
- 2. -Description of Occurrences Trojan l
On August 16, 1988 Trojan was at 100 percent power with reactor coolant loop.
"B" flow transmitter (FT) 424 out of service for calibration with its low flow bistable tripped.
Venting of FT-424 caused a transient on the instrument's high side sensing line which is common to two additional loop "B" flow transmitters (FT-425 and FT-426).
FT-425 sensed the momentary pressure transient to be a low RCS flow condition, and tripped its bistable satisfying the two-out-of-three logic required for a reactor trip on low RCS loop flow.
Prior to performing maintenance on the loop flow transmitter, a shift supervisor, control operator, and instrument and control technician discussed the calibration evolution.
It was recognized at that time that the three loop "B" transmitters shared a common high side pressure tap, and that it was necessary to valve in and vent the transmitter slowly to preclude affecting the other loop "B" flow transmitters.
However, when venting air from FT-424, the technician nonetheless induced a low pressure spike on the common leg of the transmitters. Since reactor power was above the P-8 setpoint, a reactor trip occurred.
Sequoyah Unit 2 l
On May 23, 1988 with unit 2 at 70 percent power, a reactor trip occurred from a low flow signal on RCS loop 4.
Prior to the reactor trip, the loop 4 i
channel II low flow bistable was tripped, and the transmitter was being removed from service for calibration.
Having isolated the transmitter from I
the common high side sensing line, instrument mechanics failed to follow the surveillance procedure when relieving system pressure from the line. Upon opening the instrument's high side sensing line isolation valve, a momentary reduction in pressure in the common high side line occurred which was j
sufficient to cause inadvertent actuation of the loop 4 channel III low flow bistable. This bistable, actuation completed the necessary coincidence for a reactor trip to be generated, since reactor power was above the P-8 setpoint.
ll The surveillance procedure was written to calibrate the differential pressure J
I transmitter without removing the transmitter from service and without loosing any of the RCS liquid in the sensing line. As the Sequoyah plant design incorporated an isolation valve in each of the transmitter high side legs, one of these valves was closed to isolete the channel II flow transmitter from the common high side line. The procedure then instructed the mechanics to " crack" the transmitter high side tee fitting to relieve system pressure prior to connecting the calibration equipment. Contrary to the procedure, the mechanics relieved pressure from the sensing line by momentarily opening a drain valve on the high side line. This drain valve was located at the lowest point in the sensing line.
Additionally, the drain line was routed to a
l' 1...
4, l-closed drain system which made it impossible to determine how much fluid was lost when the valve was opened. When the drain valve was opened, enough RCS water was drained out of the high side sensing line to cause a void in the sensing line. When the high side isolation valve was opened to return the transmitter to service, the void in the high side drain line caused a pressure drop in the common sensing line to the channel III transmitter, generating a reactor trip.
Salem Unit 2 On April 21, 1988 while at 100 percent power, a false RCS loop 3 low flow signal was generated with the reactor above the P-8 setpoint, cJusing a reactor trip. A maintenance technician, having repaired a leak on the low pressure side of the RCS loop 3 flow channel 11 transmitter, was returning the transmitter to service when he induced a pressure spike in the common high side sensing line. This momentarily caused the other channels to indicate a
~
false low flow condition in the RCS loop.
The pressure spike was caused by personnel error. When removing and returning the transmitter to service, the technician did not utilize procedures. The procedure for work on these transmitters identified the specific valve manipulations required to minimize pressure perturbations in the transmitter sensing lines. However, during removal from service, the technician incorrectly isdated the low pressure side of the transmitter before isolating the high side sensing line, and failed to open the equalizing valve as required by procedure.
Diablo Canyon Unit 1 On January 8,1988 with the unit at 100 percent pcwer, a reactor trip occurred due to a simulated RCS loop 3 low flow signal. A plant technician was refilling the sensing line from a reactor coolant loop 3 transmitter when he inadvertently depressurized the common high side pressure leg at a rate sufficient to produce a low reactor coolant flow signal on two out of three loop 3 flow channels.
With the plant operating above the P-8 permissive setpoint, the required logic of a reactor trip was produced.
While reinstating the flow channel to service in accordance with a surveillance test procedure (STP), the instrument supervisor and techniciar agreed to refill the transmitter sensing lines by filling from the RCS rather than by backfilling with a separate water source. When filling and flushing the sensing lines through an instrument drain valve, the technician depressurized the common leg of the flow transmitters at a rate sufficient to cause a simulated RCS loop low flow rate reactor trip.
I In discussions with the technician, NRC representatives concluded the technician j
was unaware that valve manipulation could create a sufficient perturbation in j
sensed flow to cause a reactor trip.
)
.o
-5.
l-3.
Corrective Actions by Licensees Co'rective actions are still being evaluated at the Trojan facility.
Considera-
)
e tion is being given to revising the procedures involving RCS flow transmitter J
removal from service and reinstatement to service to provide specific instruc-tions on instrument valve manifold manipulations. Additionally, the use of l
signs at the instrument valve manifold locations, warning of the potential to generate reactor trips, is being assesseo.
At the Sequoyah facility, the licensee reviewea previous reactor trips for occurrences of similar events involving common interactions. Three additional reactor trips were identified. Two of the trips were associated with temporary installations of brush recorders. The third trip on April 5, 1982, from 100 percent power, occurred when instrument mechanics were backfilling the low pressure side of a RCS flow transmitter. Momentary surges from the backfill pump caused all three transmitters to spike, generating a reactor trip signal. As a result of this trip, backfilling of RCS flow transmitters was replaced with a " wet cal" method whereby a water box is used to keep the system water solid.
As a partial corrective action at Salem, signs were placed at the RCS flow transmitter valve manifold locations to caution personnel about the sensitivity of the transmitters to manifold valve manipulations.
At Diablo Canyon, the licensee revised the transmitter reinstatement STP to include precautions and detailed guidance and controls for returning the RCS flow transmitters to service. Additionally, other STPs involving transmitters that share a common tap were reviewed.
The licensee determined two STPs associated with steam generator level instrumentation also required revision.
4.
Other Sensing Line Configurations 1
During this study, AE0D also assessed other reactor trip instrumentation i
design configurations involving shared sensing lines. Specifically, sensing i
line configuration on pressurizer pressure and level instrumentation, and steam generator level instrumentation were examined for the use of common sensing lines. On the Westinghouse plant reviewed, certain pressurizer pressure and steam generator level instrumentation were found to have common sensing lines (see Figures 2 and 3).
However, a review of the licensee event report data base from January 1,1987, through December 1988, identified no instances of inadvertent reactor trips on low pressurizer pressure or low steam generator level. This can be explained as follcws.
In order to create a low or high RCS pressure reactor trip signal, a two-out-of-four coincidence is required. Maintenance activities performed on one pressurizer pressure transmitter could affect the other transmitter sharing the common sensing line, causing an inadvertent reactor trip. However, in order for a low pressure reactor trip to occur, pressure in the common sensing line must be inadvertently reduced about 300 pounds. With the pressure transmitters being located close to their isolation valves, and generally having small sensing lines, an inadvertent reduction in pressure of this magnitude during transmitter maintenance and return to service activities is
~..
[
unlikely, considering the large inventory of steam in the pressurizer steam space and associated pressure instrument piping.
Inducing an inadvertent high pressure condition in the sensing lines is also considered unlikely, as_it is difficult to unexpectedly increase pressure in the sensing lines during maintenance.
The steam generator level instrumentation includes an arrangement whereby a wide range (WR) transmitter shares a common sensing line with one narrow range (NR) transmitter. A reactor trip signal is generated by steam generator low low leve'l from two-out-of-three NR transmitters on one-out-of-four RCS loops.
If maintenance was.being performed on the WR transmitter sharing the common leg with the FR transmitter, while another NR transmitter was out of service with 1
its bistable tripped, a perturbation might be introduced in the common leg causing an inadvertent low level reactor trip. Apparently, this combination of circumstances is rather uncommon.
5.
Safety Significance Unplanned reactor scrams cause unnecessary challenges to plant safety systems, and provide undesired opportun ties for unexpected equipment malfunctions and 1
operator errors. Operational experience data indicate that one out of five reactor scrams is accompanied by another fault, equipment failure, or human error (ref. 5). Accordingly, both the NRC and the nuclear industry have focused increased attention on the issue of unplanned reactor scrams at U.S.
plants.
In June 1984, the Nuclear Utility Management and Human Resources Committee (NUMARC) established an industry average performance goal for unplanned reactor trips, and an industry group was asked to track reactor trips, analyze causes, and recommend methods for reducing the number of unplanned reactor trips. A number of good practices were developed to help licensees to minimize the number of unplanned reactor trips. Owners groups have also established programs to address reactor trip reductions.
FINDINGS AND CONCLUSIONS The analysis and evaluation of recent inadvertent reactor trips from sensed low RCS flow at Westinghouse pressurized water reactors has resulted in the following findings and conclusions:
(1) A search of the LER data base revealed four inadvertent reactor trips occurred in 1988 at Westinghouse plants from sensed low RCS flow during maintenance of RCS flow transmitters.
(2) The inadvertent RCS low ficw trips were attributed to personnel errors, procedural inadequacy, instrumentation design configuration, and a general lack of understanding of the sensitivity of RCS flow transmitters with common sensing lines to manifold valve manipulations.
(3) Corrective actions taken by involved utilities included revising appropriate procedures to include precautions and detailed guidance for
i.
p 7
instrument manifold manipulations, incorporation of " wet cal" techniques, and the use of signs at valve manifold locations to caution personnel about transmitter sensitivity to valve manipulations.
l (4) Nol inadvertent reactor trips resulting from maintenance activities on steam generator level or pressurizer pressure instrumentation sharing common sensing lines were reported between January 1987, and December 1988.
REFERENCES 1.
Portland General Electric Company, Licensee Event Report No. 50-344/88-2f.
Trojan Nuclear Plant, dated September 15, 1988.
2.
Tennessee Valley Authority, Licensee' Event Report No. 50-328/88-24, Sequoyah Nuclear Plant Unit 2, dated June 17, 1988.
3.
Public Service Electric and Gas Company, Licensee Event Report No.
50-311/88-06, Salem Generating Station, dated May 10, 1988.
- 4.. Pacific Gas and Electric Company, Licensee Event Report No. 50-275/88-02, Diablo Canyon Unit 1, dated February 8,1988.
5.
Draft NRC NUREG 1275, " Operational Experience Feedback Report - Progress In Scram Reduction", June 1988.
m___________.__
Y_
A 5
A 1
n o
i tc S
e 5
S P
1 A
4 e
T rus e
W s
,ep ru
,r a
ss s uPT O
wep L
or a LPT F
TN
]
A L
[
O A
OC R
O T
A we C
A o
l E
F R
ggm ~ e3E a23$;gB
mu e
d P
,n l
e e
ml i
3 b
n r
D F
u b
T g
u E
P T
xH r
M E
t F
s@
D re T
s N
t P
r a
o e
A t
H k
a j
er P
e H
a Sn M
.m
e G
AteV s"
E n'
l 1
T i'
o r
I R
S u'
I n
o T
P Cy s
n 1
R i
s g
a a
M@
O T
py P
B TA R
N WM E
N W
E F
S G MD A/
/
E MEI F
ATI e
/
L ESP 3
TYM
\\
SSS N
I TF LS 8
r s~;E l g
b
g r
e s
S L
pd r
c x
ol S pd ol e
y o o Co z
c A
LC RLC iru C
s ek E
E.
sf T
T elin l
rea PRT Q
I l
M ETS Y
Spo Co RL S
TP TL F
f E
ILE e
P L 1J 4
A R
D e4at 1J rL NA e4 r
L 1 J c
R T
g L
nim E
lp m
Z ay SS I
T R
E L
U TL S
'm S
V C.
av 1
C E
P C.
R g QA
,r O
P e
,y tsy lgg g,P V
V C
C e
P rur O,
WPT h
se g
st ee s r e nyAw
,i:E Ig=
ii p3{$"