ML20032B861
| ML20032B861 | |
| Person / Time | |
|---|---|
| Site: | Big Rock Point File:Consumers Energy icon.png |
| Issue date: | 11/02/1981 |
| From: | Bordine T CONSUMERS ENERGY CO. (FORMERLY CONSUMERS POWER CO.) |
| To: | Crutchfield D Office of Nuclear Reactor Regulation |
| References | |
| REF-SSINS-6820 GL-81-18, IEB-80-14, IEB-80-17, NUDOCS 8111060457 | |
| Download: ML20032B861 (26) | |
Text
T T
CORSum8IS POW 8r
(:- - -,
- Company g,g, g k
o.n.r : ovvic : ts4s w t P <n.ii no.a. J ekson, ui 492ot. (si73 7ss o4sa
. Of
[2 November 2, 1981 L
- J clq Q N,
g,j 0
3 g
L~,
fl
'j (9
p
%r,g
,u j6-9,x;x f/
Director, Nuclear Reactor Regulation Ob Att Mr Dennis M Crutchfield, Chief Operating Reactors Branch No 5 US Nuclear Regulatory Commission Washington, DC 20555 By letter dated May 22, 1981, Consumers Power Company submitted a response to NRC Generic Letter 81-18 dated March 30, 1981 concerning the diverse scram dump tank instrument criteria. On October 2, 1981, via telephone discussions with NRC-NRR representatives Messrs Paulson and Eccleston, additional clarification was requested of Consumers Power Compauy prior to receiving exemption from the criteria for diverse instrumentation. The staff specified four areas of conern as potential for common cause failures of instrumenta-taion which should be addressed by Consumers Power Company. These areas were identified as hydrodynamic forces, human error, crud build-up and manufactur-ing errors. The staff concluded that the primary concern of adequate protection from hydrodynamic causes was satisfactorily addressed in our May 22, 1981 submittal. Additional support is provided in our submittals dated July 31, 1980 (two letters, same date, to I&E, Region III), Od. der 27, 1980 and January 26, 1981, which address the concerns of IE Bulletins 80-14 and 80-17.
However, the staff requires additional information regarding the other three areas of concera, ie, human errors, crud build-up and manufacturing errors.
This letter responds to that request.
Big Rock Point Scram Discharge System Design The design of the Big Rock Point (BRP) scram discharge system is much different from the systems used in newer BWRs.
This unique design (recognized in SER in section 3.1.2 and 3.2) consists of a single, instrumented scram l
discharge tank (SDT) rather than the separate scram discharge and instrument volumes.
Reference attached diagram to the Appendix of this letter. The SDT instruments are conneded to an upper and lower two inch instrument header that attaches directly to the SDT on the top and bottom through normally hG /
locked open manual isolation valves, not to a vent or drain line.
The water discharged from the control rods during a scram is piped through four inch j
branch headers (4) to a six inch header and finally to the top of the SDT.
The one inch vent and two inch drain lines have their own penetrations into the SDT.
[
oc1081-1022a142 111060457 011102 DR ADOCK 05000 0
(
r
^%
Dennis M Crutchfield, Chief 2
Big Rock Point Plant November 2, 1981 Human Errors I
The present functional test performed at each refueling outage involves allowing the dump tank to fill following a scram signal during shutdown. This test results in filling the dump tank and verifies the communication path between the dump tank and instrument piping as well as operation of the level switches. No valve or instrument manipulation is required to perform this test and valve line-up is identical to that required for normal power operation.
Crud Concerns Past operating experience and sampling done during the mandated scram tests in 1980 have not revealed any problems related to crud accumulation.
In 1980, following a scram at high temperature, the suspended solids levels in the scram dump tank were measured and shown to be very low.
A suspended solids level of 205 ppm was measured after the first scram and a second scram the same day indicated a suspended solids level of 13.6 ppm. These low levels are by design in that both water sources feeding the dump tank are filtered (either by the internal filters in the drives or the hydraulic pump discharge filters.) Demineralized water is used for switch calibration again limiting the potential for crud accumulation in the dump tank. As noted above, system design incorporates relatively large piping and valves insuring the minor crud accumulation which could occur at these solids levels has little potential for drain, vent or instrument piping blockage. Failures such as crud buildup, sticky pivots, spring failure, water logged float, switch binding hot shorts, switch corrosion and vibration damage are all readily detectable by our present maintenance and test program.
Manufacturing Error i
Although random failures have occurred we believe the 19 year performance of the dump tank level switches provides an adequate record of good application, design and fabrication. Big Rock Point has had approximately 131 Magnetrol float switch years of experience with five level switches in the dump tank instrument group and four other switches in the reactor water level instrument group of the reactor protection system.
(These four switches were replaced by another type of level instrumentation about 1971 for ALARA purposes, not for reasons related to reliability.) Two failures in the original reactor water level switches in 1964 were caused by mercury tilting switch binding due to tight clearance between the housing cover and the electrical wiring to the switch. One failure of the dump tank hi level alarm switch to reset and one failure of a dump tank hi level trip switch to trip in 1974 were traced to sluggish mercury tilting switch operation. Corrective measures have been taken to reduce the probability of recurrence of both types of failures.
Environmental Conditions and Random Failure The dump tank and instrument system have also been exposed to the pressures and temperatures and hydraulic effects of about 30 scrams during power operation with the primary system above atmospheric pressure as well as the hydraulic effects of another 100 scrams during primary system pressures at oc1081-1022a142
3
=
Dennis M Crutchfield, Chief 3
Big Rock Point Plant November 2, 1981 atmospheric pressure. The y radiation levels at the switches is currently at i
100 mr/hr and this is not deemed to be potential common cause concern.
Although minor vibration at the switches exists from conductive coupling with the positive displacement rod drive pumps, no random failure has been attributed to this and it is not deemed a common cause concern.
In addition to electrical fail safe design, the magnetic feature of the switch mechanism is failsafe, thus any effect causing loss of coupling between the magnet and the attraction bar would not cause loss of safety function.
The normal environment of the switches in a well ventilated room is 60-90 F with dew points generally 15*F lower and no corrosion problems have ever been noted on the switch units.
PRA Evaluation To confirm our belief that water accumulation or the dump tank does not significantly contribute to possibility of a failure to scram, a fault tree approach to evaluate the likelihood of accumulation, using probablistic risk techniques has been performed. Benefits obtained by installing diverse instrumentation on the risk associated with failure to scram are evaluated.
The analysis is attached as an appendix to this letter.
Conclusions from this analysis are as follows:
1.
Diverse level instrumentation addresses common cause failure modes primarily attributable to manufacturer error. Diverse level instruments will not eliminate common mode failures resulting from human error or crud buildup plugging the instrument lines. Other actions are required to deal with these failure modes.
2.
Water accumulation in the dumpDh-is a~'significant contributor to the failure to. scran onTp'if the SDT is not regularly drained and vented, since this presents a greater challenge to the instrument pe r f o rmance. This sitr.ation occurs only during extended periods with the drain and vent lines isolated or plugged.
3.
Divelse level instrumentation deals only with increasing the realiability of sensor inputs to the Reactor Protection System and does not consider the alternate course and benefits obtained by reducing the probability that these switches will be called upon to actuate. By itself, operation with dump tank vent and drain valves open reduces the failure to scram probability so significantly that the benefits from diverse level instrumentation are minimal.
i Based on the above discussion and conclusions reached by probabilistic risk techniques we believe the following measures are adequate in reducing the potential for common cause failure to an acceptable level of risk.
1.
Reduce or eliminate the importance of common cause failure of the dump tank level switches due to manufacturer and design by:
ocl081-1022a142
Dennis M Crutchfield, Chief 4
Big Rock Point Plant November 2, 1981 Continue to normally operate with and periodically check the a.
dump tank drain and vent valves open. This will be controlled via Tech Spec submitted for NRC approval on October 3,1980.
Approval has not been received to date.
b.
Continue to perform preventive maintenance, calibration and functional checks on existing level instrumentation during each refueling outage. Calibration and functional testing is presently required by existing Technical Specifications.
2.
Reduce the potential for common cause failure of the dump tank level switches due to human error by:
Operate with vent and drain valves open as discussed above.
a.
b.
Perform a functional test after all level switch maintenance or calibration. As discussed earlier, the functional test is performed without any valving manipulation or other activity, the system is checked as is for proper communication between the dump tank and instrument piping as well as operation of trip inputs into each protection channel by allowing the dump tank to fill following a manually initiated scram signal.
3.
Insure that the probability of crud plugging lines associated with the scram dump tank systems is small.
a.
As stated earlier, past experience and sampling has not revealed any problems related to crud accumulation. However, to reduce the probability of potential crud problems causing common mode failure, we will commit to testing at each return to power following refueling or reactor scram to verify that proper draining of the dump tank and hydraulic communication to the instrument header exists.
Consumers Power Company believes the manner in which the Big Rock Point scram dump tank is operated and maintained makes the probability of a failure to SCRAM due to water in the dump tank small.
Diversification of dump tank level switch design addresses only a portion of the causes of this type of failure and does not significantly improve the reliability of the SCRAM system.
Alternate, cost effective methods of insuring dump tank reliability exist which take the approach of reducing the demand on the level switches. The most significant of these methods insures the dump tank drain and vent lines are open and free cf obstruction.
Consumers Power Company cannot at this time justify the need for this modification. We request the NRC to examine our evaluation as well as the design'of our dump tank in detail.
i T
oc1081-1022a142 4
-n.
,-nr
~
Dannis M Crutchfield, Chief 5
Big Rock Point Plant October 29, 1981 Further justification from the NRC is also requested should you conclude that modifications to our dump tank level instrumentation system.is necessary.
This justification should include the detailed technical evaluation performed in regard to dump tank system reliability in order to establish the validity of benefits associated with the modifications.
C Thomas C Bordine Staff Licensing Engineer CC Director, Region III, USNR NRC Resident Inspector-Big Rock Point 3
oc1081-1022a142
)
I APPENDIX
-Evaluation of the failure to SCRAM potential from water accumulation in the SCRAM Dump Tank and methods of dealing with this failure.
INTRODUCTION A fault tree approach to evaluate the likelihood of a failure to SCRAM has been undertaken to address the need to replace existing level' switches on the.
ERAM Dump Tank with diverse instrumentation (design & mfgr). -Fault tree quantification of the failure to SCRAM at Big Rock Point had not been
. performed as a part of the PRA due to the complexity and redundancy of the reactor protection and associated systems. Although a small part of these systems, the Scram Dump Tank lends itself to this type of quantification due to the simplicity of its design.
This evaluation was performed as a result of a telephone conversation with the NRC on 10/02/81 in which they expressed the concern that a common mode failure of the SCRAM dump tank level switches could' lead to the inability to SCRAM without operator knowledge of the situation. Previously, the NRC, in a letter dated March 30, 1981, issued requirements for diversity of dump tank level switches. This letter required the installation of redundant and diverse level switches (with respect to both design and manufacturer) on the SDT as well as provide assurrance that common mode processes such as hydraulic forces, plugging due to CRUD -buildup, and human error were not significant contributors to level switch actuation failure. As an alternative to installing diverse instrumentation with respect to design, switches were permitted 9hich. operated on the same physical process (ie, float switches) so long as periodic testing with the reactor at power were performed.
On May 22, 1981, CPCo responded to the Generic Letter concluding that diversification of the dump tank level switches was not necessary at Big Rock-Each of the common mode concerns expressed by the NRC was addressed in this letter and the unlikelihood of these failure modes was generally attributed to the unique design of the dump tank (i'e, good coupling between the tank and instruments, low hydraulic forces on SCRAM, low particulate concentration in SCRAM water).
4 i
i oc1081-1022a142
!I.
I
r 3
2 Description of Problem The main concern is fai'.ure to insert control rods during a SCRAll due to an accumulation of water in the Scram Dump Tank. The attached fault tree accounts for the accumulation of water in the SDT by two methods:
a.
Undrained water from previous SCRAMS and tests.
b.
Leakage into the tank from external sources; leaking scram valves and demineralized water.
The tree concludes 1.
Failure to SCRAM as a result of high water from previous scrams and tests requires the following.o occur simultaneously:
a.
Plugging of vent line or drain line or failure closed of either of the valves in these lines.
b.
Failure of all level switches in the SDT (indication of an empty tank when in fact it is full) c.
SCRAM initiation 2.
Failure to SCRAM as a result of water accumulation from external sources required the following simultaneously:
a.
Water addition to the SDT from leakage SCRAM valves or demin water valves.
b.
Insufficient drainage of the water addition c.
Failure of all level switches in the SDT d.
SCRAM initiation Evaluation of Fault Tree The fault tree was evaluated for two cases, power operation with the dump tank drain and vent valves open and operation with these valves clcsed. The position of these valves is important in that the rate of water addition to the dump tank must be significantly greater for the case with the vent and drain valves open than with these valves closed in order to accumulate sufficient water to prevent reactor SCRAM. The position of the valves therefore affects the probability that water collects in t'e SDT.
Assumption made in evaluating the tree include:
a.
Closure of either of the valves or plugging of the drain or vent line are sufficient to prevent dL.ining of the tank.
b.
Closure of either VRD-21 or 22 is sufficient to prevent communication of water in the SDT with the level switches.
oc1081-1022a142 4
4 e
s
3 c.
Leakage from a single scram valve or from the demin water system is not sufficient to overcome the drain capacity of the SDT.
d.
Position of the SDT CV's (drain and vent) is checked each month.
e.
SDT level switch annunciation is checked each shift.
f.
Proper communication of SDT with level instruments is checked at each refueling outage (s annually).
g.
Proper communication of drain and vent with sump is checked at each refueling outage (* annually).
h.
Proper operation of level switches is checked at each refueling outage (s annually).
1.
The dump tank will fill with water if the drain and vent valves are closed.
General results of the fault tree follow:
Failure to SCRAM given drain and vent valves normally open per year:
P 9 pen
=S
[V + VP+(LS, )(LSad)]
[PP3 + PP2 12a 33, 03
+ CV
+CV
+
+F
- ( 11b
- 12b} cl D
Scram Failure of level switches Probability of accumulation Probabili ty for random & ccmmon of water in dump tank and per yr.
cause reasons.
failure to drain Failure to SCRAM given drain and vent valves normally closed per year:
P Closed
=S
.0 }
Scram Failure of level Assumes dump tank Probability switches for random will fill if drain &
per year
& common cause reasons vent valves closed Symbol definition is attached to the fault tree.
The total probability to SCRAM due to a high dump tank level is determined by multiplying the above equations by their respective fraction of times with the vent and drain valves open or closed.
Failure probabilities for each of the components in the fault tree was taken from Appendix III of the Big Rock Point PRA.
Plant specific data was used for a particular component where available. Othe rwis e, generic data was used as a basis for the failure rate.
Common cause failure of the ievel switches due co human error and plugging of the line to the instruments was explicitly modeled in the tree.
Common cause failure due to like manufacturer and design was generally modeled with a p factor method using 10% coupling (ie, 10% of all failures were considered to be a result of common causes). This is consistent oc1081-1022a142
4 with the treatment of common cause failure modes used in the Big Rock Point PRA.
Numerical results of the evaluation are attached to the fault tree. Several plant " modifications"to deal with common mode failure of the level switches are evaluated as well (including diverse instrumentation). The numerical results lead to the following conclusions:
1.
Diverse level instrumentation addresses only those common cause failure modes attributable to such things as manufacturer and design.
Diverse level instruments will not eliminate common mode failures resulting from human error or crud buildup plugging the instrument lines. Other solutions are required to deal with these failure modes.
2.
Water accumulation in the dump tank is a significant contributor to the failure to SCRAM only if the SDT is not regularly drained and vented. This situation occurs only during extended periods with the decin and vent lines isolated or plugged.
(The NRC's letter of March 30, 1981 deals only with increasing the reliability of the level switches and instrument volume piping and does not consider the alternate course of reducing the probability that these switches will be called upon to actuate).
From the fault tree it is believed the following course of action will to insure water accumulation in the SCRAM dump tank is not a significant contributor to the probability of a failure to SCRAM.
l 1.
Insure the probability of crud plugging lines associated with the SCRAM dump tank, particularly the drata and vent lines, is small.
(Whether or not diverse level instrumentation is installed).
2.
Eliminate er reduce the importance of common cause failures of the dump tank level switches hie to manufacture and design by any of the following (in decreasing order of effectiveness):
Operate with the drain and vent open.
a.
b.
Install diverse dump tank level switches with respect to manufacturer and principle of operation.
c.
Provide a continuous monitor of dump tank level which is periodically checked by operations.
3.
Eliminate or reduce the importance of common cause failures of the dump tank level switches due to human error by ar.y of the following (in decreasing order of effectiveness):
Operate with the drain and vent open a.
b.
Perform a functional test af ter all level switch maintenance c.
Provide position indication or position locks for VRD 21 and 22.
oc1081-1022a142
5 CPCo should commit to performing only those actions which make the likelihood of a failure to SCRAM due to water in the-SDT low when compared to the probability of a failure to SCRAM frcm other causes. The probability of a failure to SCRAM was presented in our submittal of March 1,,j981 ~ concerning RPT ano in Appendix VII of the. Big Rock Point PRA (3.5 x 10 per demand).
The effectiveness of several of the preceding recommendations is evaluated and attached to the fault tree. Assuming CPCo a) eliminates plugging of the SDT piping as a likely event (an argument to this effect was presented previously to the NRC in our May 22, 1981 letter), b) cormits to normally operating with the drain and vent valves open the failure to SCRAM probability resulting from dump tank water accumulation is as follows:
P = (8752) (<<10-7) + ( 8* ) (9.2 x 10)
(8760)
(8760)
~7
= 8.4 x 10 per demand or less than 3% of the currently assumed failure to SCRAM probability.
Cost Benefit Analysis The cost of diversification of the scram dump level instrumentation is uncertain.
In developing the design and cost of the new instrumentation consideration must be given to the following:
Equipment environmental qualification requirements.
a.
b.
The need for additonal penetrations in the tank.
c.
The degree of fail safeness to be implemented in the design.
d.
The extremes of operating conditions which occur in the dump tank.
e.
The manner in which the new instruments must interface with the reactor protection system.
f.
The physical process by which the new instruments must operate.
None of these considerations has becn examined in detail at this time.
As a result only a range of cost estimates can be : ad in performing a cost beneift analysis.
C (3.5 x 10-5) x
=
(8.4 x 10-7)(2.7 x 10-5)(1.0)(59.4)(10 )(18)
Cost-benefit ($/ man-rem) x
=
- Assumes "normall3 open" SDT valves are closed no more than 8 hrs /yr during power operation (consistant with current operating practice).
In fact the amount of time the valves can be closed is dependent on the rate of water' leakage into the SDT for which this analysis took no credit.
1
,.~
..~
L 6
Cost of modificatioc ($)
C
=
i 7'
8.4 x 10
=
Failure to scram per demand due to dump tank water accumulation
-5 3.5 x 10
=
Failure to scram per demand-from all causes (App IV PRA).
-5 2.7 x 10
=
BRP ATWS core damage probability per year (App I PRA) 1.0
=
Containment fa'. lure probability given ATWS core damage.
4 59.4 Latent fatalities resulting from release category
=
BRP-3 (Table 1.2-PRA) 4 Man-rem per latent fatality (NUREG-0739)-
10
=
a
]
18
=
Years remaining in operating license.
Assuming the cost of this modification will range between $20,000 ar.d $200,000 the cost benefit is $3,000 to $30,000 per man-rem.
This analysis assumed instrument diversification will eliminate entirely'the
' probability of a failure to scram as a result of water accumulation in the dump tank.
In fact this is not the case. Common cause failure of identical instrumentation in each RPS channel can still occur _and the small probability of instrument failure due to human error cannot be addressed by this modification.
4 The modifications proposed in our October 16, 1981 submittal to the NRC are
]
estimated to range in cost benefit between $600 to $10,000 per man-rem with the majority costing less than $3,000 per man-rem.
Using these modifications
)
as guidance, diversification of the dump tank level instruments has marginal cost-benefit.
i i
f i
i I
1 i
i ic1081-2139a123 3
i
-7 i
SYMBOLS j
BASIC EVENTS i
AN
- SDT h1 level annunciator failure CVC
- CVNC11 (Vent Valve) failed closed 11 CVC
- CVNCl2 (Drain Valve) failed closed HV
- Demin isolation valve leak j
- LSRD08A Switch failure g
- LSRD08B Switch failure B
l LS
- LSRD08C Switch failure C
l LS
- LSRD08D Switch failure D
- LSRD08E Switch failure E
MS
- Flow to scram dump tank greater than drainage capacity (Multiple SCRAM valve leakage)
MS
- Flow to SDT from one'or more SCRAM valves 2
0
- Operator failure to drain dump tank after last SCRAM or test 1
0
- Operator failure to recognize. drain and vent valve 2
position indicator closed 0
- Operatar failure to respond to hi-dump tank alarm OHV
- Failure to close demin isolation valve after last test PP
- Vent line plugged 1
PP
- Drain line plugged 2
PR
- Air line to vent and drain valves ruptured 1
P R,,
- Air line from Master Scram valves to SDT drain and vent SV's ruptured S
- SCRAM probability per year SVD
- Solenoid valves to SDT vent and drain failed closed V
- Valves RD-21 or RD-22 isolated ic1081-2139a123 i
I 8
- Fill or vent piping to level switches plugged TOP EVENTS CV
- CVNC11 closed after refueling surveillance 333 CV
- CVNC'.2 closed after refueling surveillance 12a CV
- CVNC11 closed during power operation Hb CV
- CVNC12 closed during power operation 12b F
- Flow into SDT with drain valve closed C
F
- Flow into SDT with drain valve open D
- SDT level switches failure to trip (LSRD08 A,B,C,-D) ad LS
- Failure to SCRAM reactor on high dump tank level (operator or switch failure) ic1081-2139a123
l 9
FAULT TREE ALGORITID!S All Cases LS
- (
a) (L3 ) * (Ls ) (LS )
ad b
c d
LS,
= AN + LS
+0 3 F
= HV + OHV + MS 2 CV
=0
[SVD + PRI + PR2 + CVC 2]
12a CV
=02 [SVD + PR + PR2 + CVCg7]
g 7
Drain Valve Open 12b 12a lib 11a F
= "5 d
1 Drain Valve Closed CV
= 1.0 g
CV
= 1.0 11b F
= 0.0 d
Drain Valve Open =
F Open = S [V + VP + (LS, )(LSad)l 1+
2+
12a 11a
+0
+Fd + (CV11b
10 COMPONENT RANDCM FAILURE PROBABILT]
AN
- SDT hi level annunciator failure
-6
-6 1.4 x 10 /hr (8 hr) = 5.6 x 10 2
assumes:
- 1) Annunciator failare is primarily result of open circuit.
-6 1.4 x 10 /hr Ref: Table III 4c Cable, open ckt.
- 2) Shift check of annunciator ekt CVC
- CVNC11 (Vent Valve) failed closed g
1.01 x 10 /hr (740/2 hr) = 3.73 x 10~
-6
- 1) Diaphragm open valve FTRO 1.01 x 10 /hr Ref: Table III-3 Item 16
- 2) Valve position checked each month CVC
- CVNC12 (Drain Valve) failed closed g
~0
~
1.01 x 10 /hr (740/2) = 3.73 x 10 '
- 1) Same as CVC; HV
- Demin isolation valve leak 1.1 x 10~ /hr (8760 hr) = 4.82 x 10~
2
- 1) Hand valve FTC 1.1 x 10~ /hr assumed to be same as leak Table III-Aa
- 2) Exercized 1/yr and SDT leakage checked / month LSA
- LSRD08A Switch failure 2 x 10 /hr (8760 hr) 8.76 x 10~
2
~0
- 1) Level switch FTO Ref III-4a 2 x 10 /hr
- 2) Tested annually LS
~ ^** 8 b,c,d a
MS)
- Multiple SCRAM valve leakage (drain valve open) ic1081-2139a123
T 11
[(7.6'x 10~ /hr) (8760))3 ; 1.11 x 10~
2
- 1) Assume at least 3 valves required to leak sufficiently to overcome drain capability
-7 2) 7.6 x 10 /hr Air valve fail TRC Ref III-3, Item 16
- 3) Annual exercize and maintenance MS
- Single Scram valve leakage 2
1.0
- 1) Assumes a high likeli 1 of small leakage to the dump tank O
- Operator fails to drain. ump tank after last test or SCRAM g
2 x 10~ /d)(.1/d)s = 2 x 10~
- 1) Action is familiar, procedure available, location familiar Ref: Table III-8 (low or moderate stress) one operator - complete dependence.
0
- Operator fails to reccanize drain and vent position indicator closed 2
(.1/d)5 = 10-5
- 1) Passive inspection Ref Table III-7, Item 6
- 2) Checked once per month (s10 checks / year) 0
- Operator failes to respond to high dump tank alarm 3
([(2.2 x 10-3) 19 +1] 2.2 x 10-3) (,3)S00 20
- 1) Failare to respond to annunciator 1 out of 1
~
Table III-7 Item 7 and 8 1.1 x 10 *
- 2) Low stress level (normal op), 2 operators zero or low level of dependency
- 3) Failure to read annunciated lamp each shift (annunciated display-passive examinabtion)
.1 Table III 7 Item 6 OHV
- Failure to close demin isolation valve after last test 1 x 10~ /d
- 1) Ref: Table III-8, Item 3 ic1081-2139a123
12 PP
- Vent (drain) line plugged 7
PP N 0 2
- 1) Low particulates in dump taak water
- 2) Large piping associated with dump tank system PR
- Air line rupture g
-6 R
(8 x 10~ /ft-yr)( 1_)(100 ft) = 8 x 10 2
_ 10 2
1) e same as pipe rupture Ref Table XII.11 Assp/f t-yr, monthly check on valve position indicator, 8 x 10 N100 ft air line est.
S
- SCRAMS per year s2/yr
- 1) Ref: Table XII.4 SVD
- Solenoid valves to SDT failed closed
-7
{(1.5 x 10 /hr)f20 hr)] = 2.92 x10~
2
- 1) Both SV's required to be closed to close vent and drain valves
-7 2) 1.5 x 10 /hr SV FTRO Ref: Table III-4a
- 3) CV position checked every 'nonth V
- VRD-21 or VRD-22 closed af ter maintenance
-6 (1 x 10" )(3 x 10~ ) (.51) = 1.53 x 10
- 1) Failure to open valve per maintenance procedure 1x 10~ /d Ref: Table III-8, Item 3
- 2) Failure to note valves are in wrong position CRD Check list Ref: Table III-7, Item 30 3 x 10-3,9 j
- 3) 50-50 TR-32 performed af ter level switch maintenance and failure to perform TR-32 =.01 Ref: Table III-7, Item 5
(.50 +.01 =.51) 4)
If TR-32 always performed af ter level switch maintenance ic1081-2139a123
(.
13
~
V = (1 x 10 )(3 x 20" )(.01) = 3 x 10' VP
- Level switch fill or vent line plugged NO
- 1) Same as PR I
l l
i i
ic1081-2139a123
14 COMPONENT COMMON CAUSE FAILURE PROBABILITY Level Switch Failure LS
- AN + LS + 0, anr e
s
= 5.6x 10-6 +.9 (8.76 x 10-3) = 7.88 : 10-3 LS
(*
(*
adr (LS,
)(LSadr) # LS
= (7.o3 x 10' )(1.24 x 10) +.1 (8.76 x 10' )
u
= 8.76 x 10
Instrument Piping Valve Position Error V =.9 (1. x 10' )(2.9 x 10-3)(.51)
V
=.1 (1 x 10' ) (.51)
CC V=V
+V r
cc (2.61 x 10-6 + 1 x 10)(.51) or (2.61 x 10-6 + 1 x 10)(.01)
=
-5
-6
= 5 25 x 10 1.03 x 10 if TR-32 performed after switch maint.
Crud Blockage of Instrur.ent and Drain and Vent Lines 0 as crud plugging drain vent and instrument lines is considered N
low probability.
ic1081-2139a123
I 15 QUANTIFICATION OF FAULT TREE an_d EVALUATION OF T10DIFICATIONS
- 1) Baseline P
- (8 ad)l 1+
2 11a
- open an CV12a + 0 + MS +CV11b +
12b
= ?{5.25 x 10-5 + 0 + 8.76 x 10 ~'][0 + 2(3.73 x 10~9) +
2 x 10"
+ 1.11 x 10"
+ 2(3.73 x 10~9)]
= << 10~ /yr
+
+ (LSan ( ad)][1.0]
closed
= 2[5.25 x 10-5 + 0 + 8.76 x 10-4)
-3
= 1.86 x 10 /yr
- 2) Perform dump tank test af ter level switch maintenance (Eliminate Human Error)
~0
~
P
= 2{1.03 x 10
+ 8.76 x 10 ']
elosed
-3
= 1.75 x 10 jy,
- 3) Perform dump tank test af ter level switch maintenance and add diverse instrumentation (Eliminate common cause switch failure)
-6 P
= 2[1.03 x 10 /yr + (1.53 x 10" )(3.76 x 10-3))
elosed
-6
= 2.37 x 10 /yr (Assumes diverse instrumentation reduces failure rates to random failure probability) ic1081-2139a123
(
JL 16.
U Nyu I
l W
y
+
it M
1 l
g4 i=
4 1
t 1
1 le 1.
2
' t 4
[1 b
E 5
s
=
c I
I f
\\t v
S 8
s!
u
-a N
N N
D w
D b
h v
a 9
i C
i o
z A
g 4
5 W
j C
=
'Q,
I I
b bS NA e
E 8
a QP 2
C o Eo i
ba c
%@c e
-t C
3 N
E N
N m
Ed
?
did aw>
h
@o.
=E o
i yo s U=-kN t
di 30 i
h@~
X c
+
e A@
FAL"./I IREE 17.
ruelu,e ASCRAM D ue T. 4,9,
1
% p h kLeeI O
I I
SDT Level So.hb 5, lores O
5 Lue/$wohks ll,g f, $ D 1" RD*b Level De Mo* Wep ew l
l l
' Io. lure. +c Dman!
Ficaa Ioic $DT SD T Mer GrtG k' T h n t.as h f Il
% Cof em em CV/2A l cve28 l
l Fn
__... l 5Dr Venf CV NCl2.
cvNCit op fa,/,<. r, ricus n io
% n}eSOr e r Dea,,,
c Dra.n )
(Ven j }
AHeap& ro sur d*%
- a. h. "L)m,,,
L,nes Pf&
Closed Closeg Drom SOT Dmon Vake Cpen valg Chwi,
/-
+99 2 l
cvNCil Op 4 lure. To Q,fs 7a" Nolt %s. bd-E"3 '* Y'A 4f;,pfc ggy f j,g Open 4 Dd ale.br OM" t.as kage.
O2 D
Mb e em l
CVN(. it
& r,i3 A shk Cheed Aur To CVHCll cyc,,
f FAULT TREE 18.
(Trsnsfer Point 2.1)
CV NC # 2 2.I \\
Cloxd O
I I
Cwc t2 Op [
%.9 mees Ind 4 Carls,to 4 /s /o Pe;nte coen Og h
r%
I I
NC [nsh Elf CV HC12 Slak M ZZ\\
cy c <3 A
CVCl2 em I
I kJONf&
Wo Air kers klf (Ind SVdC 22 D PR, I
A., oee p.m sygncg 3 Maskr::en:m qg ato ha
-s svo peg
I 19' FAULT TEEE (Transfer Point 3.1)
Claa mlo SUT 3'I
~
#h D ***
Valve Cosed 1
I I
sc vent e<
,,,,,,u "Drsm WNes SDT Clos ed
\\
b em l cvs2%
cvilB l
l l
CVC NC 12 cvt NC al Flem from Row Gem re, led Clesed Nied Gesed Mm,n vkke one. er Mere 5Y5km Summ %ks i
M5 2 L%
l "Demm. Iso, Cp Ala< k
't volge g y close isc.
Valva. r NV OHV
(
~?AULT TREE 20.
(Transfer Point L.1)
Level 5=3 W S Do dof l
- I '\\
Trop &
c.
l m
I i
i VRD 21 or L5 RDOS 'S Lent Lokh 2g closed Dodd Pesul?
p;iterVer+
in 2 Tro D Lor,c Plu44cel
~
\\
v, 0v On l
L54 a 09 Doed 45eDoS i
4+ /mlo/c A-D Co 4f 35 Yeo o Ac+va+L (3
A I
Cp k Mt SDr Hi level A""'ned h r Egspcnd +o a none.a k e XeoIoee C3 m-I Annuncnks LSRDC8 E 5, lure.
ra,/ ort.
Ou O~
t
.r.
e I
FAULT TREE 21.
(Transfer Point 5.1)
LSAv f-p m ola r t em I
i CMnnel I C hannt.12 4, lore faulve s f3 O
i l
I L5RDct A l LSEDCS S ls2DG8C LSRDc8 D To ulure Q,jur c.
Q/or e foulure O
O O
O LS4 LS r3 LSe L5 p g
..