ML20054L579
| ML20054L579 | |
| Person / Time | |
|---|---|
| Site: | Byron, Braidwood, Zion, 05000000 |
| Issue date: | 06/25/1982 |
| From: | Hanan N, May R, Singer B SCIENCE APPLICATIONS INTERNATIONAL CORP. (FORMERLY |
| To: | |
| Shared Package | |
| ML20054L576 | List: |
| References | |
| NUDOCS 8207080262 | |
| Download: ML20054L579 (48) | |
Text
- _ _ _ _ _
l D
D D
PROBABILISTIC EVALUATION OF HIGH PRESSURE LIQUID CHALLEliGES TO SAFETY / RELIEF VALVES Iii THE ZION, BYRON / BRAID'..'00D PWR PLANTS b
D Sub-i tted to:
C0!ODNWEALTH EDISON COMPANY o
Reliability & Design Engineering Group 9
Pecort Prnaared by i;. A. hanar R S. Na; g
Technical Contributions oy:
I S. Singer O
Principal investigator R. M. Crawforc O
Scienct Applications, in;.
On brook, Illinois June 25, 1982 0
8207000262 820701 PDR ADOCK 05000295 P
.O TABLE OF CONTENTS PAGE q) 1 1.
INTRODUCTION.....................................................
1 O
2.
EVENT TREE ANALYSIS..............................................
4 2.1 General Discussion..........................................
4 C) 2.2 Extended High Pressure Injection Event......................
5 2.3 Col d Ov e rpres s u ri z a ti on Event...............................
9 2.4 Main Feedwater Pipe Rupture Event...........................
12 C) 3.
F AU L T T RE E AN ALY S I S..............................................
16 3.1 Power Operated Relief Valves (PORVs) for Zion Plant..................................................
16 C3 3.1.1 At Power.............................................
16 3.1.2 Miscalibration.......................................
23 3.1.3 Cold Shutdown........................................
26 3.2 Modifications to PORY Fault Trees for C)
By ro n - B r a i dwo o d P l a n t s......................................
30 3.3 Failure to Recover from Spurious Safety 38 Injection...................................................
C) 4.
S U M ' AR Y O F R E S U L T E...............................................
40 40 4.1 Zion Plant..................................................
40 4.2 By ro n - B r a i d w o o d P l a n t s.....................................
0 4.3 General Conclusions.........................................
40 REFERENCES.......................................................
43 O
t
.O i
O LIST OF FIGURES PAGE
()
2.1 Centrifugal Charging Pump Characteristic Curves......................................................
7 C) 2.2 Event Tree for Extended High Pressure Injection atPower....................................................
8 2.3 Event Tree for Cold Overpressurization......................
11 2.4 Byron /Braidwood Stations:
Pressurizer Pressure, Water O
Volume and Pressurizer Relief for Main Feedline Break with Offsite Power (Fig. 15.2 B/B FSAR).................
13 2.5 Event T ree f or Mai n Feedl i ne Brea k..........................
15 3.1 Power Operated Relief Valve Sketch with Actuation O
Air Supply.................................................
17 3.2 Zicn Overpressurization Detection Instrumentation...........
18 3.3 Zion Power Operated Relief Valve PCV456 Actuation Circuit (Typical)..........................................
19 0
3.4 Reactor at Power - Fault Tree for Unavailability of Both PORVs - Zion........................................
20 3.5 Probabil i ty T ree Diagr am f or Calibration Task...............
25 1.6 Reactor at Colc Shutcown - Fault Tree for
)
Unasailability of Both PORVs - Zion.........................
27 3.7 Reactor at Power - Fault Tree for Unava ilabil i ty of Both PORVs - By ron/ Bra i owood..............
31 3.E Re :t r at Cold Snutocen - Fault iree f or C)
Unava ilaoii i ty of Both PORVs - Byron / brai cwooc..............
34 3.9 by ron/ Brai cwood - PORV s Logi c Di agrar........................
37 O
O O
ii
.O LIST OF TABLES
- O PAGE l
3.1 Failure Data for Fault Tree Analysi s........................
24 4.1 Summary of Results for Zion Plant Frequency g
of Liquid Di scharge f rom Safety / Relief Valves...............
41 4.2 Summary of Results for Byron /Braidwood Plants Frequency of Liquid Discharge from Safety / Relief 42 valves......................................................
O O
O O
O O
O O
iii
O 1.
INTR 3UCTION 4
1
- g Based upon the requirer.ents of NUREG-0737, caners of nuclear power plants must perform plant-specific evaluations to ensure that the Power Operated Relief Valves (PORVs) and spring-loaded Safety / Relief Valves (S/RVs) are i
operable and provide effective pressure relief under the possible range of g
discharge conditions.
The Electrical Power Research Institute (EPRI) has 2
conducted the PWR Safety and Relief Valve Test Program to provide a generic basis for addressing these requirements.
This program is nearly completed, and data is available to provide the starting point for evaluations by O
individual utilities.
EPRI experimental results, as well as various independent analyses *, have shown that saturated liquid and subcooled liquid discharge through the safety
-O valves constitute the most severe challenges to such valves and associated piping networks.
There are two major reasons for concern about S/RV liquid dischargc.
- O First. S/RVs open rather quickly, with stroke times of 50 ms or less, so that the discharge piping may experience large dynamic loads as a wave front of liquid entors pipe segments.
PORVs open more slowly over periods of 1/2s to 1s, and thus give rise te smaller dynamic loads.
Further, the loads are
.O proportional to mass flow rate, whicn can be up to 6 times higher for liquid flow compared to steam flow.
Thus, from the point of view of discharge piping stresses. the combination of licuid discharge through an S/P.V represents the greatest challenge.
O r;
Secona, it has been snown by EPRI test results and by incependent analyses ~
Inat piants witn long SRV inlet piping may be subject to chatter oscillations lO of the spring 1:edec valve.
Sucn oscillatory behavior is most likely for succoolea licuid discnarge, because tne high mass flow rate generates water-nanmer pressure waves of very large amplitude.
Oscillatory behavior may also be observed curing the expulsion of succooled licuid 1000 seals.
'O each of these two concerns, the situation is most severe f or f ar-succooled i
For l
liquid distnarge, which gives the maximum mass flow rates.
Saturated or
[
10 1
l
O
slightly-wbcoeled" liquic discharge is screwhat less trouniesone, because there may be suf ficient fleshing at the valve so that two-phase flow effects con sutstantially reduce the mass flow rate and thus mitigate the event.
It is believed that liquid discharge can be tolerated both from the perspectives of discharge piping stresses and oi valve oscillations.
- However, piping nalysis is under w y t verify the ability of discharge piping to O
accomodate the required loads.
High pressure liquid discharge, while admittedly a conceivable event, is extremely unlikely when considered in the context of available systems, operating procedures and time for operator ction, and infrequency of the initiating events.
It is desirable to O
understand the frequency of occurrence of SRV liquid discharge in order 1) to provide perspective and recognition of the relative unimportance of subcooled discharge, in view of relatively high stresses which may be computed for such lO discharges, and 2) to provide a more rational basis for defining realistic inlet conditions for SRV discharge by eliminating those conditions which are shown to be very unlikely.
6 O
The present study uses techniques from Probablistic Risk Assessment to evaluate the frequency at which S/RV liquid discharge may be encountered in the Zion, Byron, and Braidwood plants of Commonwealth Edison Company.
SAI has performed similar analyses to evaluate liquid discharge risks in Boiling Water 7
O Reactors.
Baseo upon previous generic analysis by Westinghouse, and upon additional plant-specific hand calculations, event trees are oeveloped to oualitatively describe the event secuences which may cause S/RV liquid discndrge anc to ioentify tne system functions anc ooerator actions which may O
f avorabiy or unf avorably aff ect tne outcome.
Faul. tree analysis is then used to ouantitatively evaluate the faiiure probabilities tur tne reouirec system Extensive use of previous researcn results,9,W,H 8
anc operator responses.
f or even*. initiation f reouences component failure rates, and human error O
probabilities are usea nnere applicabie.
Resuits snow that liquid discharge from the pressurizer safety-reliev valves is extremely unlikely for tht Zion, Eyron and braicviood plants. Subcooled O
liquid discnarge cey occur at ins rate of 9.6E-8 events / reactor unit-year at O
2
1 i
19 1
i
]
j Zion and 5.8E-8 events / reactor-year et Eyron/Braidviood.
Saturated liquid l
j discharge ray occur with frequences of 3.65E-8 and 1.05E-8 events / reactor-year lG at Zicn and Byron /Braidwood respectively.
1 l
o
!G 1
1 I
t i
IO i
I t
't I
i l
1 i
1 1
1 l$
I f
l l
s 4
O t
O 3
O 2.
EVENT TREE ANALYSIS 2.1 General Discussion O
in support of the EPRI/PWR Safety and Relief Valve Test Program, Westinghouse has performed a generic evaluation of the expected range of fluid in7et ccnditions for pressurizer safety and relief valves for plants designed by Q
7 Westinghouse. The resulting report provides a comprehensive discussion of all transients with the potential for high pressure discharge as well as bounding calculations for the actual conditions to be expected by 2-loop, O
3-loop and 4-loop plants.
That report provides the starting point for our analysis.
However, as a generic bounding analysis the Westinghouse report quite properly 1) asssumes O
multiple system f ailures without evaluating their likelihood, 2) ignores the helpful effects of any operator action, and 3) fails to take the credit for plant-specific characteristics which mitigate the events.
The present report
- 1) modifies the important Westinghouse reported sequences in accordance with O
the plant-specific characteristics of Zion and of Byron-Braidwood and 2) evaluates the frequency of the important sequence by incorporating results of fault tree analysis described in Section 3.
O Tne casic conditions f or liquid discharge (of any kind) is that the pressurizer pressure reach the S/RV set point of 2485 psig and that the pressurizer. eater level rises to tne too at the same time.
Because of the different leveis of interest in tnese scenarios. as discussed in Section I we O
will distinauisn netweer.:
i) oDening discnarge of subcooled licuid ii) opening distnarge of saturated licuic iii) opening discnarge of saturatet steam followed oy delayec discharge of saturated licuid.
Tne Westincncuse report identificc tne transients potentially leadina to O
licuid distnarge as:
O 4
O ii FIAT, events -
a)
Fu d.sater Pipe Rupture
- b)
Accicental Depressurization ii)
Exterded High Pressure Injection Event (Spurious Safety injection) v, iii) Cold Overpressurization events - a) mass input event b) heat input event The star-ed events will be analyzed in detail in this report.
FSARs for
.n I2 13 Zion and for Byron /Broidwood make it clear that the Feedeater Pipe Rupture is the only FSAR event of concern for these plants.
The " heat input event" for cold overpressurization was included under the mass input event, because o
the latter is,1) more likely to occur, 2) much easier to characterize quantitatively (without extensive system transient analysis), and 3) less easily mitigated because it is a very fast acting event.
O 2.2 Extended High Pressure Injection at Power Spurious actuatien of the safety injection system can be caused by operator error or by a f alse actuating signal.
Should the operator fail indefinitely
<O to recover from safety injection and in particular fail to trip the centrifugai charging pumps, the primary system may fill with subcooled water.
Following an initial drop in pressure due to primary coolant contraction, the system would begin to repressurize after the pressurizer becomes solid aue to O
continuec cnarging pump operation.
Tnis event has a fairly hign frequency of 0
occurrence, but it is also very easy to cetect and terminate.
Generic data for PWRs leacs to a frecuency of 1.6 X 10-1 events / reactor-year f or Byron-Braiasood, which is a relatively new cesign anc tnus representative of C
tne general PWR oooulatior..
Pian: soecific cata ootained for Zion dicta ws a frecuency of 6.0 X 10-for tne Zion pian..
A Safety Injection Signal (SIS) results ir. e reactor trip followed oy a o
's turd 1ne trip.
The ietoown is automatically isolated and is tnerefore unavailable t or pressure relief.
The centrifugal charging pumps force ECCS water into two primery colc legs.
Since there is no letdown (which in any case coes not nave suf ficier.: capacity for mitigation) the primary loop water n'
ira sntory steadily increeses.
At first, the cressure croos cue to the coolant O
5
O contraction, until the prenurizer water icvel increases to the top.
With pressurizer control lost, the system would then pressurize until high pressure liquid is discharged thrcugh the PORVs or safety relief valves, unless there
.O is appropriate operator action or successful discharge through PORVs.
For Zion and Byron /Braidwood, successful operation of only one PORV is sufficient to remove liquid supplied by both charging pumps and thus to
.O eliminate the possibility of SRV liquid discharge.
Figure 2.1 shows the t
characteristic curves for the charging pumps at Zion and Byron /Braidwood. At 2335 psig and saturated steam conditions, the single-PORV flow rates O
comparable to these curves are 58.3 lb/s for Zion and 457 gpm for Byron /Braidacod. For saturated liquid PORV discharge, mass flow rates are even higher.
It is apparent that one PORV provides adequate pressure relief for extended safety injection.
The first branch of the event tree of figure 2.2
- O reflects this fact.
In the fault tree evaluation of PORV, only automatic actuation is considered; no credit is taken for operator action of the PORVs.
t Given that both PORVs fail, a simple mass balance shows that at least 20 O
minutes is required for tne pressurizer bubble to coliapse and for liquid discharge to occur.
However, there are clear cut operating procedures for recovery from safety injection which will require the operator to reset the 515 within a few minutes.
Furtner, this is an event which is neither O
extremely rare nor dif ficult to interpret, so there is a high likelihood that the event will be successfully terminated by the operator.
The human response is analyzed in Section 3.3.
The comouted human error orcbabilities are expected to be very conservative, since the 20 minute response time obtained O
f rom the simoie mass naiance is exoectec to be a very snort and tnus conservative estimate of response time, which may be on tne order of nours.
Wnen tne nut.ericai results are obtainea f rom section 3, it is found tnat tne M
f reauency for SRV liquid distnarge following a sourious safety injection is 3.6E-E events / reactor-year f or Zion and 9.9E-9 events / reactor-year for By ron/ Bra i c.-;occ.
l
'O Furtner, the S/RVs would in any case be first chalienged by steam discnarge followed by a transition to saturatec or sligntly subcooled liquid disch9rge, lO 6
. x
,.e' O
,N
- .z -
O
- co -
a.
Zion
- l C G C --
4 O
m S.
s.a ac
?,
1200 -
m s.a oc L.
m
- 17. !!!IJ TLC 2 F E D-T 'O U
C Es7 E i, UC AL C'. NC I NO T':.".?S a
J tC3 -
l'UC 20.
40 00 80 100 220 Q
T1.0.; RAT E - LE!!S EC (CCO c O
~
w!,:u';* a ; e t.1; ??p0.!*1':=cr.:
o Cf =i6 sf G1 C ast... Ptw
^
(;
f L*
1s;' twist r
e..,.
4 0
b.
Byron /Braidwood E
j l'
~
C I t 7 8 I f * I '. Castti=
P.**
8 r
(;2g!.!!: f(
I;-t 1
P!Af; t*
ws;-:
,O wn a
30 c. 1-e e.t:i. u: nt3 i::3
/
g b
sc :,
c.
1 t
i I
C c
c_~..
c 1 sO f r. ~,.......:.,n...:,
t.......
ric.. c.1 o,.
C Ef;TRI FU^22.'
CHARG:::.^ PU"? CHARACTERIETIC CURVES O
7 i
l
.O SPURIOUS 0iE PORY OPERATOR S/RV O
SAFETY AVAILABLE C0!1TROLS LIQUID It? JECT 10Ti SAFETY DISCHARGE AT POWER If?]ECTI0ti
!O rio lO
,0 6.0E-1" b (1.6E-1) o i
14 0 t
c.4E-4
- lg (5.6E-4) l 1.lE 4 3.6E-8 9 9E-9 (1.lE-4)
YES O
^Zior.
O UByron /braiawoo:
Fig. 2.2 EVEtiT TREE FOR EXTEiDED HIGH PFESSURE li?JECTIO!i AT POWER g
l O 8
O reducir.g the potential for chatter instability as well as the anplitude of dynamic loads en discharge piping.
O 2.3 Cold Overpressurization Event A cold o.erpressurization event represents the most likely source for liquid S/RV discharge for which the liquid may not be preceded by any steam O
discharge.
The entire system begins from cold shutdown conditions, so that 1) subcooled liquid is present throughout the primary loop so that S/RV discharge will assume maximum mass flow rates and thus face the greatest problems with respect to waterhammer instability and downstream piping loads, and 2)
O pressure control is inadequate to prevent rather rapid pressure excursions from occurring, since no steam bubble exists in the pressurizer.
An verpr ssurzation event from cold shutdown conditions can be mitigated by O
enhanced mass input to the primary loop.
Only the centrifugal charging pumps are capable of raising the primary loop pressure as far as 2485 psig for S/RV discharge.
The initiators essentially involve spurious and uncontrolled O
pening f a fl w p th througn the centrifugal charging pumps.
The mest credible events causing cold overpressurization are:
a)
Failurt of tne Air Suoply System in the Zion Plant would cause o
tne charging flow control valve and the letdown valve to fail closed.
This in turn causes a net injection of mass by the centrifugal charging pumo and a very high rate of primary loop pressurization.
A value of 8 X 10 failures / demand was O
used f or tnis analysis; this numoer is tnought to be very large arc tnus extremely conservative.
[This scenario aces not not apply to Byron /Braicacoc; a loss O
of instrument air pressure would aiso close tne valves (CV8324A and CV8324B) upstream of tne Regenerative Heat Exchanger.
Therefore, there would be no mass addition to the RCS in Byron /Braiov;ood.]
O O
9
O b) Failure of the ch?rging fia.. control valve to operate as required; this could be caused by local valve failure or local failure in the air supply to the flow control valve.
Based g
upon data from reference 11, the rate at which this initator occurs was determined to be 1.2 X 10-3 failures / year.
Since the reactor is at low pressure and the centrifugal charging pumps fl a rate increase with decreasing RCS pressure, a large mass lC injection causing a high rate of pressurization would occur.
Although the letdown path is available throughout the transient the letdown relief rate is inadequate to effectively mitigate the event, and no credit is taken for letdown in the probabilistic O
analysis.
Further, no credit is taken for rapid operator action, because very high pressurization rates (up to 100 psi /s)7 have been predicted O
for such events.
For Zion, the PORV set points during cold shutdown mode are set to 500 psig, O
as long as the PORY control mode Selector Switch is correctly placed at the
" AUTO LOW TEMP POSITION."
Zion operating procedure GOP-2 dictates this switch placement so that procedure violation would be necessary for the set points to remain at tneir at-power position.
O For Byron /Braidwood, the coerator must correctly re-set two swittnes (PORV Control Selector Switch to "AUT0" and cold overpressure control switch to"0N")
in oroer to ensure inat tne correct PORV setooint is cqosen.
O Tne first branch ir, the event tree in 2.3 involves failure of the coerator to follow procedures for mode selection.
According to NUREG/CR-1278, the human error probability (HEP) for failure to follow this type of procedure is.01.
O The second brancn concerns an operator error of turning on a second centrif ugal charging oump in violation of proceoure GOP-2.
'O The PORY f ailure rates cepena uocr. wnetner tne PORVs have ceen set ir the colc shutdown mode (correct) or in the at-power moce (incorrect).
In the latter O
10
O COLD TOM CC',3 DL EENTRIFUS;L ONE FORV S/RV OVERPRESSUEIZATIO:i ICDE SELECTOR CHARGI:;5 PU:P #2 AVAILABLE LIQUID SWITCH It; 0FF DlSCHARGE 1
CORRECT POSITIO!i
.O f40 0.99 (0.99) 4.3E-5 8.4E-8 O
(4.3E-5)
(5.lE-8) 4 YES i
0.99 (0.99)
.O
- ti0 1.0E-2
~
(1.0E-2)
O 4.3E-5 8.5E-10 (4.3E-5)
(5.lE-10)
YES a
2.0E-3
)
(1.2E-3)b
!O 14 0 0.99 (0.99)
O j
j i
5.4E-4 1.lE-8 i
t l
8 (5.6E-4) ygg (6.7E-9) 1.0E-2 (1.0E-2!
!O l
-- tio l
1.0E-2 (1.GE-Z'
- O l
l 5.4E-4 1.lE-10 i
(5.6E-4)
(6.7E-ll)
YES i
" Zion
- O bByron /Braicwood Fig. 2.3 EVEriT TREE FOR COLD OVERPRESSURIZATION lo 11
_,,_m-..,_,.-.,,,,-,,.,,,,,,,,,n
,an_-.w.,,,,,,v,,,,,.-.,,.n-
.,,,,,,,-~,---,e,-
O case the failure rates are identical to those developed for the at-pc.ser case (see Section 3.1.1).
Ho ever, if the FORVs are in the correct mode, then failure rates are those developed in Section 3.1.3.
g g
When nurnerical results are obtained from Section 3, it is found that the frequency for SRV discharge due to a cold overpressurization event is 9.6E-8 events / reactor year for Zion and 5.8E-8 events / reactor year for g
Byron /Braidwood.
2.4 Main Feedwater Pipe Rupture Event O
A nain f eedwater pipe rupture, if large enough, can prevent the addition of I
sufficier.t feedwater into the steam generators to sustain shell-side fluid i nvento ry.
Should the large break occur between the check valve and the steam O
generator, the water can quickly discharge througn the break causing a rapid loss of heat sink in the affected loop.
The FSAR transient response for Byron /Braidwood i, included as Figure 2.4.
O Tnis event is not analyzed in the Zion FSAR, but the results are not expected to be significantly dif ferent.
Following the injection of cold ECCS water, the pressurizer pressure and level both initially drop due to the negative surge rates caused by primary loop water shrinkage.
After about 5 minutes, O
the pressure again rises due to reducec neat removal through the steam genera tors.
Eventually saturated steam is discharged through the S/RVs at about 7 minutes into the transient followed by a transition to saturated 11ouid cischarge after 13 minutes into tne even..
Tne safety relief valves O
tnus serve as a neat (ano rass; sint to stabilize tne transient until tneir final closure at 20 minutes.
In the FSAR, no credit is taken for tne operation of the PORVs. which would O
open at 2335 psig and distnarpe saturateo steam f or pressure relief.
As discusse' in Section 2.2 by Fioure 2.1, one PORV would be adecuate to balance tne cold-water input from both safety injecticn pumos anc thus control tne primary ico;. pressure to 2335 psig, se tnat tne spring-loadec safety valves O
would never c o e r.. It is possibie, nowever, tnat tne PORvs may eventually O
12
O
_lll[lil{
l !ll}lll
_, _ y.
l IiI...!j l llll!lif l
E 2500 1
wc 2 00 3
.q o
mw 2300 C
L 2200 ew N
2100 J
c I
Dmm 2000 wc' 1900 10
,3 c
O w
2 Ng o
1 Ew D38 0
!$
- R Cn w
q H
g c.
S
-10 2000 lJ g
[
1750 w
3 1500 o
O c*
1250 5
t=
1000. -
O r_.
t V.
E 750 -
=_
r) l l l Ililll l l lllill:
I l l llill:
I llIllll 500 4
2 3
10 2
5 10' 2
5 10 2
5 10 2
5 10 0
PRESSURIZER PRESSURE. WATER VOLU.E BYRON / BRAID'..'00D STATIONS:
AND PRESSURIZER RELIEF FOR KU:, FEEDLINE BREAK WITH OFF5ITE POWER (Fic. 15-E.4 - S/B FSAR)
O 13 l
O triemselves discharge sm saturatec liquid af ter a transition frim saturated stean discharge.
O f;o credit is taken for operator action in the very simple event tree of Figure 2.5, although shutting down the charging pumps would at any time effectively terminate the pressure up-transient leading to high pressure discharge.
14 Operating procedures exist which require such actions to be taken once the O
pressurizer pressure has stabilized and begun its increasing trend.
The initiation frequency of this event is very small, because it involves a large break in a relatively short stretch of piping between the check valve g
and steam generator.
Based upon the WASH-1400 pipe failure probability of 1.0E-6/yr-section, the initiation frequency is chosen to be 10-6 events / reactor-yea r.
O The PORV system failure probability is computed in Section 3.1 to be 5.4E-4 and 5.6E-4 for Zion and Byron /braidwood, respectively, resulting in a very small event frequency which would be even smaller should operator action be g
included. Compared to the other events of Sections 2.2 and 2.3, the feedwater line break is thus insignificant as an initiator for high pressure liquid discharge through safety relief valves.
Further, such liquid discharge is always preceded by saturated steam discharge, so that chatter instability is O
precludec, and the transient ioaos to discharge piping would be substantially mitigated.
O O
i O
O 14
s y
1 1%1ti FEEDLif4E 0:iF. PORY S/RV
)
BREAK +
AVAILABLE LIQUID DISCK4RGE
.1 t
4 1
- -c
?.
J tio
~
)
/
l
)
(l.0E-6)b
~
a l.0E-6 r
j
-is v
l l
~5.4E-4 E.4E-10 5.6E-10 l
- {5.6E 4)
_ YES i
u
)
_. ~
Z1 ori
-~,
o 3
Byron /Braiawooc
~
Brea k be t. ween checi...'. /a and SG
~
3 Fig. 2.5
~
EVENTTREEFORMAltiFEEDLIrlEBREdF
- ~
i D
4
+V y
m/ w p.
3 s
15
.y-v,--,,.m.mm.,__..
,%g,,,m..,_
,g,,,,__
9 3.
Fli l UlE A'.TU 5:5 The analysis of the e.ent trees dsscribed in the previous section shcas that 8
the Power Operated Relief Valves (PORVs) constitute the only equipment capable of eliminating the possible liquid challenges to the Safety / Relief Valves (S/RVs) in the Zion and Cyron/Braidwood (B/B) plants.
Fault tree analysis techniques were used to quantify the unavailability of the FORVs for the Zion and Byron /3reidwood plants.
Tote, from the event trees that the availability of one PORV is sufficient for mitigation of the posible incidents.
The Zion plant is discussed in detail, and for Byrori/Braidwood only the system differences and modifications are discussed.
J 3.1 Fault Tree Analvsis of PORV's for Zion Two pressurizer power operated relief valves (PORVs) exist in each unit of the 2)
Zion plant.
Figure 3.1 shows a sketch of one of those PCRVs with its actuation air supply.
The PCRVs are actuated on a signal from the Overpressurication Detection Instrumentation (Fig. 3.2) through its actuation c rcuit shown
'n fig. 3.3.
b directec by operating proceaures, "the
~
nu operator adjusts the position of the PORV Control Mode Selector Switch according to the Reactor made of operat. ion.
If the Reactor is operating at power the switch is on "AUT0", anc if the Reactor is in Cold Shutdown rode the switch is on " AUTO-LOW TEMP."
As seen in Fig. 3.2, these different switch g
positions aictate cif ferent PORV actuation signals.
Therefore, the unavailability of tne PORVs is accendent uoan the made af operation of the keact or Tne follawing sections ciscuss tne fault-tree analysis for tne g
Reactor "A: Fower" anc " Cold Snutacen" moces of operatlor..
3.1.1 Reactor at Power 1
'O A fault tre..
r f aiiure of notn PORvs to coer. wnen tne reactor is at power
+
Was Constructed and Guantifled.
Ine fault tree is snoWn in Fig. 3.4.
hote that in Fig. 3.4 tne aetails are given only f or one of the PORVs; tne other is identica. It tne first one.
Tne results of this analysis snow an overall
,0 r;2cian unan11abilit, (for cotn POR'. s ; equal tc 5. E-4 / cer enc.
O Jg
}
A C C U t iU L / *T O f?
\\Y 3
CGE TO O P E rd tut 5:
i 3
7 f_
ru s TR u r$r n r 4
' p_{[
- A g f;
/
Ts jf
/,
O W
V rm\\
-A th' 'triu ti ei i h,~,c y
PE L I5 uru'd E A e-ts
- J P
,k.^
/
/'
FevJ E n O P E r-AT E.,
O nt.tra r v otc.
='
F v '
O FGWER OPERATED RELIEF VALVE SKETCH WITF ACTUATION AIR SUPPLV O
O 17
O O
t; O
(.)
U U
C>
u 8
U PT PT PT DT PT ri 4a3 gn5 t r, y
- 4sa, 4 *,7 tagar 2
,s
%s l
+
r I
]
/*
- s.[
l
[
1 r
l 1
des p'-%
(
j ga43 " ;..z, y, ~, t s swe s.
_,.., _,,,7
?!
g ' ' '. l '.' '. A c
a s
4a+
e,a
- {j s-i *
'I 33ea tu e..p t.
... ~.a
- 'IO ertsos g FkE4 briii) 2 3 i s-
)gg3
'-N*
!*%9 2i e 1
- c..
T
' ' ~ ~
E"sl
!.^'5]
IG'555 f'.5i_li i
I e
[. ~ g p.,],
! P o. y w a Lv F. C o ae T
- 01.**
u,.,._.,
7 i
I l
l
=
i 4 45
[
7~'tfy'
... u a t u q ".'f S
ij creo i
s.
~u st.
...t.,
_. ( e.r T..,
q yy crus f...
. _ _ _ _ _ Q_<...,<.__ _. - _ _ _. _<.....
( O F. E w i T C 64 )
- y 1} (CNE&*eTcea)\\
g
+
se r>s
- t ti t t v s.tv s c r u '. r v a u n saeq
( c.
- g u p, P(V 45%C Lew T e **'s Os ts PR * %%u s t ? ATicu D P CV 4
- G.
a s... y wer ~ sta,,ct )
"o ce t.o i....,,,y g Fig. 3.2
- 7. ION O'/EPPRESSURIZATIO"i DETECTION INSTRUMENTATION e.
e.
g
- .V::
- , 3 T o.
D F E :.
g.
es
{_ Ji __.
l
- )
'.ri L
A
/s J
O 3
.. _C1 C 5_
. b' C-R O
~ Aum-
~ L O TE *1P
--sJ AUTO
\\'
V s/
N/
N/
N/
s/
N/
C O
- e v.n ty f4 st a
J o
r rti. 5 (,CK to o
da
- PC 4 OSCX' 3
i 4
T l
t
^
/x
/
/s
'l' C $
- )
O y
s
( AtJ TO i
e
+
2.
U (1
m 3
g, N'
(J VALVE S CW'- S c
- Le
_, LS
/ L e v' '
- / 5 c t. v '-
r-1 ;
v-Ic i
s w n c H :,.,. !
.TC m
4 a._
J 3
},
t p
ryg, 3,3
,sJ ZION PO'..'ER OPERATED RELIEF VALVE PCV45f ACTUATIO:. CIRCUIT (TYPICAL) n~J
\\
i
%J Q
aghh4M--_4,__
4 ai.
__ ___ Age 2,m
_A..g%
g,,m,,_,,j
,au_a._s a
2_m
__ma a u as a we,aAm aa s&5'd_ sha.h
=h_a
.-_A..,
,,ha_m_.
,4eg au.
A_
4
- _pgp, a
g r
i M
J
-m s
l 2
l 1i b-jI5) w-m IE l.
E ere ir de I
Y.,
8 r.
1 (A..
8 i
f, -
=a d
- 9
?
O e
m N
i.
a
]
s s
Y a
L 3 *.
[1
- t
+
- g f
Om e
i r *'
E.
f CL
$5 g
y m
j 4
1 O
.i, to n
(
3 L._
e6 i.
o e
i' i$
E 1.
~
n2 4
.e m
l 8
w M;f 4,
~
-.J a.
- a..r.
W m
D M CD
=
=e l.
l 0
J j
Ch
- -a 1
w >
ia
<w D
I i'.
I e
O O
L s
1 N... b.
i W
w j*%
M
'M h:"
n
!T
{r F
r L<
L k.
k I
i i.;, i.
b y
9 i;; p;N.e eo
- t..
- r. -
c
'1 P-*
I I
W O
~
O l:
,I w
h le -
t C
.-9 l
i,.:1 8...
-O 6
be9 b
a-$
s l' ? s.? ?
0" l E '.,
- i i
O O
20
j G
G S
9 9
9 9
9 9
9 9
i i
s i
1 i
1 i
I
...,,. m...,ss.:
C*wf*?L
[
SIE=pt
.. te.1
\\
,,f
[da L
)
i l
I
.a...m. s,..,, s ww.
l,.s,,.'.,t,,..-
..,e,,.'.,t,....
.. m.s.
=
e
,,.,...,2 1
iUfj
[5WDJ
($U IUD m
(3 U
U U
7;i3E;;r
.s
~~ ~ ~ ~ ~} I l
C retsse.t c e==tt er it rpH s Pf44' 'Dit5
[_,_
R:d r-l n)
[1 )r 1
1
(
..,.,., - r,... i t s
....g,t.....,.-
_ - - -. i s,.i,,,. i t s 3.,,,t,..tst e;,.ui
.a..i i
I o ey.1 ITiF 1 if01110
[LriD10 U
()
U O
l f
f fig. 3.4 (Continued)
REACTOR AT POWER - FAULT TREE FOR. UNAVAILABILITY OF BOTH PORVs - ZION i
t i
I h
O O
O 9
0 9
9 9
9 9
O l
l l
l l
l l
tesi e. eie u.t.
i i:.:..si i
0*
(UD ci.
qL I
I FF.,. 2
- ..... s i.
.i.u...,..,-...e,.. :... "
. i t s. e.c.e.. c,
t.
.r i e..e., - c.a., v e
.6 c..
=
e.
u.
t e s.i's :
lClD
[IldO L"@"4 n U
U U
l$
g
, a.. r t.,hm. e -
v.t, 8.,..i...:.
a.-
- i...i-p;st.
l t.
LfG']
LCD
- a.,
...,.......c
... a.,c.,.-, c.7
- a... --..,.
..n...,....
,a..
.s.
e
.rt
. i. e....t
- e..f it.
3 ; e.g.
.u s,, e.g.
Lt as at t i
- L_I I ; .~ d' l
( k.,i}
I',si?C]
Gi,i*']
((i t *'
l i.
[i h. -
1 t
(.. I)- -
U C
U C
j~i;i;~;#. ;$... I'. F -
' iu r.
e.,i s....c..,~~
.e :.
t t.e ac t n.:a npa
( ')
O'
,...,. a. ~..,. #n i
,.1.,
..u,,.,
,.. i..:
i...... a u...,
.u 1
I i
nua nea
(-)
U t
I Fig. 3.4 (Continued)
J REACTOR AT POWER - FAULT TREE FOR UNAVAILABILITY OF BOTH PORVs - ZION 4
I i
i
!C)
The rain contributor to the unavail.Lility of both PORVs is a ccrmon made miscalibration of two or r. ore tcmparators ir the pressurizer pressure control j g) systen.
This failure accounts for more than 97 of the total unavailability, and it is discussed in the next section.
{
t i
The feilure data used for this analysis is shown in Table 3.1.
1
!C) 3.1.2 Miscalibration The probability of miscalibrating two or more comparators which actuate the j
l()
signal to open the PORVs is determined using technqiues described in 10 NUREG/CR-1278 These techniques for the " Direct Estimation of Conditional l
at Probaailities" were developed by the Nuclear Regulatory Commission for evaluation of human reliability in reactor operations.
The evaulation is done
! C) in detail by considering both small and large miscalibrations.
A large change is defined as one that is so extreme so as to be not normally expected, while j
a small cnange is one that can be expected to occur occasionally because of i.
variations in equipment or other conditions.
!C)
To check the calibration the tecnnician must first set up the test equipment.
I-An error in this initial setup is the initiating event for miscalibration.
Figure 3.5 presents the Probability Tree Diagram for this calibration task.
13 It is necessary to point out nere tnat tne checking of the calibration of all pressure channel comparators is done by the same technician once per refueling j
s hu t dowr..
C)
From Figure 3.5. it is seen tnat tne probability of a large miscalibration of two or more comoarators (F,) is eaual to 5.0E-6/act, the probability of a small miscalibration of two or more comourators (F,) is 5.0E-4/act, and the probability of a small or large miscalibration (F, + F ) is equal to 5.05E-4/
7 2
()
act.
The following comments are necessary for a better unoerstanding of tne Probabilit_y Tree Diagram in Fic. 3.E :
O 23
O O
O O
kLE3.1 O
O O
O O
O j
FAllHRE DATA FOR F AULT TREE AtlALYSIS f
I I
FAILURE FAILURE RATE EXPOSURE TIME j
)
COMPONEllT T10DE (1/hr)
(hr)
UNAVAILABILITY REF j
Air Operated Valve railure to Open on Domand 2.0E-3/d LER f
a Air Operated Valvo Leakage 2.0E-7 4380 8.BE-4 LER a
l Motor Operated Valve Plugged 6.0E-8 4380 2.6E-4 LER Chect Valve External Leakage 5.0E-8 4380 2.2E-4 LER f
8 Check Valve Reverse Leakage 7.0E-7 4380 3.lE-3 LER Check Valve Fails to Open j
on Demand 1.0E-4/d LER W
Solenoid Operated rails on Demand i
Valvo 1.0E-3/d WASH-1400 Pipe (Q< 3in.)
Leakage or 1.0E-9/hr/
1/3E-5/
Rupture section 4380 section WASH-140D i
I Airumulator Low Pressuro in Accumula tor 1.0E-6 Zion PS5 Bistable (includes
[
Bistable & Logic l
Relays)
Fails on Demand 6.7E-6/d Zion P55 l
i 1
Transmitter (Includes Sensor Fails to Provide b
& Transmitter)
Proper Output 1.66E-6 4
6.6E-6 Zion PS5 i
i
" Assumes test every year bMean time to detection for these transmitters (Zion PSS) i
O
'O a =.99 A =.01
.5 6 =.5
=
a O
O B' =.1 b = 0.
B = 1. 0 b' =.9
'O C' =.01 c =.9 C =.1 c ' =. 99 2"
F) = 5.0E-4 O
A - FAILURE TO SET UP TEST EQUIPMENT CORRECTLY O
- Small Miscalibration of Test Equipment a
B - For a Small Miscalibration Failure to Detect Miscalibration for First Setroint C - For a Small Miscalibration Failure to Detect Miscalibration for Second Setpoint e - Large Miscalibration of Test Eauipment B'- For a Large Miscalibration Failure to Detect Miscalibration for First Setpoint c'- For a Large Miscalibration f ailure to Detect Miscalibration for Second Setpoint Fig. 3.E g
PROBAdILITY TREE DI AGRAM FOR CALIBRATION TASE O
.O 25
~
-Q 1.
The ccmplete notatior for the ccnditicnal probabilities events is not employed but should be understood, e.g.,
is O
written instead of al A, i.e. probability of a "given A."
2.
As suggested by NUREG/CR-1278, it is estimated that a miscalibration would be equally likely to result in a large change or in a small change. This assumption is conservative since the total probability O
(i.e. the summation of the probabilities of small and large miscalibration) is used in this analysis.
A more realistic analysis would include only the large miscalibration, because the miscalibretion error will cause a PORV failure O
(prior to an S/RV challenge at 2485 psig) only if the setpoint is calibrated to a value greater than 2485 psig.
The differences between calibrations at 2485 and 2335 should certainly be considered a large error, iO 3.
It is conservatively assumed that if the technician does not detect the instrument error by the time he calibrates the second setpoint,100% of the time he will continue the erroneous calibration through the third and subsequent O
setpoints.
3.1.3 Reactor in Cold Shutdown Mode
- O There are several differences in PORV operations auring at Power and Cold Shutdown modes of operations.
Since these differences affect the PORV failure probabilities, they are discussed belov :
M i)
Pressurizer Pressure Control Signal:
In tne cold shutdown mode, as can be seen in Fig. 3.3, the PORV control mode selector switcn is in tne position " AUTO LOW TEMP" and tne corresponding actuation circuit nas a signal f rom oniy one pressure comoarator.
Tnis modification to tne tree given in Fig. 3.4 is presented in Fig. 3.6.
ii)
Miscalibration:
With tne reactor in Cold Shutcown mode tne
.O Tnerefore, oniv a verv PORVs are set to oper. at 500 psic.
ierge miscalibration will cause tne actuation of tne S/RVs,
..ncse setpoint is at 2435 csig, before the actuation of the PORVs.
O 26
. _. _. _ _. _ __m O
O_
O O
O O
O O
O O
O i
a f
l i
i I
l e - '. r e. s.*
- i, ' e r
s i CI'J I
r..._ _ _
_t i
ee r.ase
.-.. n s v a re*= re esss.:
ee
.en, te e-e i
N fliU W'
?
N I
()
\\
.d1[
.,., a,m.,., <,....m,,.,a.,.,,,,,,,,
,,.... r.,.b..,,,.w
- n. r #...,,.n.n:
- ~nwr..nn.. 5 r; a:-c...m,,v n:o m.o esaa eco euwu emue u.c a
/\\
/.i U
U U
U U
Fig. 3.6 REACTOR AT COLD SHUIDOWil - FAULT TREE FOR UNAVAILABILITY OF BOTH PORVs - ZI0fl h
f 1
i
)
i
)
{
l t
~
O O
O O
O O
O O
O O
O i
i i
i PRES 3URilER i'RESSURE CCHIRCL SICrint
]
rnituPE
[c13]
l 4
cin I
e l
I I
I l _ ___
rt;E S %R! ("ntmfi N15t0LIPoqTICN -
FUSE fl FAILS CPEN FUSE F2 F AIL 5 Cf'EN eruns r nit s tenwru C DE t
(5NJ EINE.}
[TIO El R
U O
O l
I 1 ~
_. _ _. I m
____....__L
_~
~
nistootr rnit s t enn<.rsi t t En r nt st i
fil; f f'tfi l
l t
(. _ _.r. _ _ )
I r__t ws t r j i cuosn' l
(
)
(
)
~,
u
[
t I
l Fig. 3.6 (Continued)
[
REACTOR AT COLD SilUIDOWil - FAULT TREE FOR UtlAVAILABILITY OF BOTH PORVs - ZI0t1 t
[
I t
j i
[
t b
1 4
l t:I!!i l
i
[l;(
- f (l
,[:ii?
t
- [l
,s N _
- ilIl O
I Z
rR I P
}
i E,
{
s V
vL
},
t 0 u.
R t r O
anU rEl t
P I
]
l tiLr dt i _
F U
T 5R E5 0
r '4 B
iV PC F
P O
Y T
I L
I B
A L _
T I
L A
L r
!A V
v
- 3 ru CR A
E
)
N
]
Ss 1
ET E JE d
U s
a V X C e
Ru L E D.
f h
R -
l v d
A r
nt n
O V - A r
E i
F E
r C:
tSL t
T CL n
E s
EI o
E i
HA C
R o
CF N
(
T L
I d
6 T.
s Ec L
Ri fO 3
U 1
Un G.
A n~
Si l
SU F
C EN k
g PU P
i PC C
F
[
t uP N
C W
C L
R IE O
VL D
F t CR T
f E
]
lV Y0
}
EVE L
LT VEC i
H 1
O C
'P; S
rn P
LRO i
rL j
1 oU v
V O
D
[(
C c
LN E
L t
U KSL RC O
CL lC C
E!
Rn
?
Hn L
T tv Cr 2n O
tt Vu A
N E
CP HD E
]
R Cu O
CTE E'
T n
])
VrC r
3 1
n r
i A
LEP C
Oi
~
v I.
V - n Et C
E I
E P
[
KSL R
bs CL lf El l v Hn oi Cr fo 2
~
v v
~
r
_r
_J L_ o uA_
I
. vs lL
?[
1 vn
[
O. r E
H C
2 E, v"
C9
[
E F~
EVf 9~
V E t.
LRD i
n t
7 v - n T
E
[
tSL tL il H0
~
Cf
~
lIiI:!!
l ) jil!j i
]'
fji!i
,tl
,i
i 9
'O i
Lased on this, and using the pecbability tree diegram in Fig. 3.5, j
4
)
the miscalibration error wes taken to be 5.0E-6/act, correspcnding l
t t
2 to a large r.iscalibration only, o
iii) Air Supply From Instrument Air System:
As discussed in Section i
2.3, the failure of this system is one of the initiators for a i
r l
Cold Overpressurzation in the Zion plant.
Therefore, it is assurted that the instrument air system fails, so that the j
j loss of air supply for the PORVs is represented in the fault l
l tree (Fig. 3.6) simply by failure of the air accumulator or failure of check valves and piping.
t
- O l
The results of this analysis show an overall median unavailability equal to l
4.3E-5/ demand for the cold shutdown mode.
The main contributors are f
i combinations of single failures in both valves.
l
?O i
3.2 Modifications to PORV Fault Trees for Byron-Braidwood Plants I
The fault trees for failure of botn PORVs to open when the reactor is at Power O
and on Cold Shutdown modes of operation for Byron-Braidwood plants are shown in Figures 3.7 and 3.8, respectively.
The only differences between the Byron /
I Braidwood and Zion plants are:
1 O
i) wnen tne reactor is operating at power, the PORV controi mode selector switch is required to be on "AUT0" (see Fig. 3.9).
l However, there is a probability that the operator leaves that switcn on "CLOSE" anc this failure, by itself will result l
O in an unavailability of tnat PORV.
By comparison, in tne Zion olant tne PORV control moce selector switcn ooes not nave position 'CLOSE" as seen in Fig. 3.3.
(
ii)
Wnen tne reactor is in Cold Snutcown Mooe of coeration, tne
.O acciaent tnat mign leac to liuuia challenge to tne S/RVs is a Cold Overoressurization Event.
As discussed in Section 2.3, i
the initiating events for Zion includes a failure of the air supply system, anc thus this event is not present in the n"
fault tree for tne ODening of the PCRVs (Fig. 3.6;.
- However, for tne B/L olents a f ailure of tne air suppi,y system is not r
30
~
._._m
_A n.
m,um--ma,,a.
y a
p,..
Ja.h.-m am a me 4m_awwm-h_A..e.---.We
@m-4_g u A
R a
w d. mLJs h 2, 2A.mW
&A_ me_..
JE 4, s.4 s.l,_m_.,_A____-.-_h_
me. mu h_-_-.-as--%_,
.w..a 4
.O
...a4 I
.a t
13.
6 t
i La a
+
9 W.
e ~. b 1
l$
W 5
uw&
O e
O
.9 I
C4 N
O Oo
- p
- me..M.
,tt,,i o
'j v
J 9
is.
<x a.
)
I 40 i
"a' N
S,==
O%
,l N
i cc i
WD
,9
=.=,
l
~
m.e 3.A a
g l
e[f ce
==
n
.o f
- ?T
$6
-. v *j O
1 b
M'd Q-
,.a bP Fb g y1 v.
w
.e y
H o
O-l cc
_a W
G e
Er o
k i
=*
r.m.s v e a, 4
H i'
E d
E o.
-J i
4
=
~
n cc
..e u
4 v*
m
~
a l
_l 7..'.
- r y~
. a 7.,
Q (h
- 0..*
~
m d
l.
o l
e s>
b U
lb 3
i Ls.
cr O
t 6
t,a.!
l i' l
W G
cr w
i...
H i
lv
. S,, m
-J d A.
l
\\
J
.?.
~h
'h-v' l
)
.a
- 6. 6 Ior
~
O L.-
sN 0
e.
t W
O ow
}
t k
L.-
1 i
W fe I a E
I t,
../
O.
,.1 I '" 7 f.u- \\
C 9
j :. i' c
L
.' 5 j
.o e-i r
I l
t..
.i.
t e
i ;..
.. o
.U :.
u. " ?_
l tc.,
s e
i.,..3 ls N g a *; e
..i=
9 i
31
..- ~..
.. ~. - -... -
f Y
I t
?
i i
t i
I i
i l
l i
,etssaeirte ents bet
(?=,t**6 S1 Chat 88LL'8!
b,,.
ma l
~
T l
,,,..,'.,t,....
,ani.L
,,.a. s. e c.s s.., <,
.,,..t,,,, a..
..t.,
- m.,. s,,, -
s u
.m
.e,
- i. ai m
mn essa amo ruu
~
t (n)
U U
U U
i 1._.___
i 6
,'!556#? ( ** * *t t ret 35;st t%ew.tg j
rtasa f elt s P(ust f aits H.rD tim i
t-o t <..1
-..,-,t_..s, 1
1 i
,,3,
,a,.its i. 3.,ir, r ot s, e n i...a
- i. s i v,1, cart ?
rutret I
6 x3J r%BiL*J (NifD CFh'n l
tO U
U O
Fig. 3.7 (Continued)
REACTOR AT POWER - FAULT TREE FOR Uf1 AVAILABILITY OF BOTH PORVs - BYR0ft/BRAIDWOOD I
o
O w
O O
O O
O O
O O
t e,, c.....r..s,... t.
T- ~
iw
(@
,. ~,.u. -......
er.........,.
n.. t.e...cr c.
..... s,..e.
. it.,. e. t. s
.< r,-
c....
.t i
.a IdiI UiR J t" tic' 13 g
U U
t
....r..,
.r... v..s,.. ~.
...e
. t.. i.. s,s,,.
ww u ;
" ;a
{t i
{
'I.
__....I.'..
... : L.;.
i
.._.___.L t
[ e.-
... i u,.?.,.,. c.
i.... t s. :.s.
rs, n.s,r..;c,
._. r.:
i..
r.-t t... t.w.e.e.
- c. t..,:...(..t *i es t t.-s.t..,.s,.i r.
r t, vi
,e e.
.....s
....i se ei e:t j
Wii 1 ' A'f.!
Ik%J (2'19D nii!U U5:!J,
I~'
Il I )
U C
O C
k._)
1 l1 1
--t.,..
i..
l,.
.m..
t t.s P.
n..,
v.e t
)
( )
I f[t-i I
l
-.._.....l.f..t.o.,,
t ** t D...
t * ?..t t
....,.rt e t,. t. 2
.*t
.. s...
,,i...t
..e I
t,..i a.;r. i
()
()
i Fig. 3.7 (Continued) i i
REACTOR AT POWER - FAULT TREE FOR U'lAVAILABILITY OF BOTH. PORVs - BYR0t1/BRAIDWOOD
[
t i
--__.~_e-,-
1 e
o e
e e
e e
e e
e e
1 i
E e
l l
l f
1 6
l l
{
I l
I e-,..:. 3 r..t,e 1
t'! u j
I l
c2m Ir l
B
_L__
j i
l__ ____,t
- 45 %e 7 8 l t 1
_ Free Ptveita fp1<S l
P!'y 75 e*(w 1" P'f m LA)
I(3h EIE) b
/3
(\\ \\)
,,I s...1.,..........
1
,,.. s s..., m.., -
.,.. =.,,.,<,
..., =.,. L,,.,
m s,
..___t,.,...
e....-..
i
......,..t....s...
- e.. e,..
.m.....,,,,,.
...,.e>
im i-o uuua ssma esm et.n 3
/\\
/_\\
U U
U U
U n
I l
J l
l Fig. 3.8 i-l REACTOR AT COLD Slit lTDOW!! - FAULT TREE FOR UtlAVAILABILITY OF BOT I
i I
l'
?
i i
-.n-
1 i
i 1
I 4
1 I
6 i
rpESSU'l?En amES3Uag CONI?OL SICNDL r n lli nE j
\\_.
i i
G
^;-
l
[
i I
l i
_ _ ___ L
, FUSE F 2 T D,iLS Crt N
___.L._._...
t FrEsiUpE consr.rt e t M nL 199n t ico -
FUSE fl FAILS D'EN I
I l
1
[
l r(+ios rniLs
( "a m": r*00E i
i
_e i
?
um....
eiss em w
(d) l g*
I
[
I k
PI".in4LE I Pit s t orso,n g ligg r ot sr j
FUtrPT l
l
. -. l F fjiffi_T~J l
I Ffiiyi-l l
t
[
r i
Tig. 3.8 (Continued) j PrActoR Al col n qitrinow;l - FAllLT TREE FOR UilAVAILABILITY OF BOTH PORVs - BYR0ti/BRAIDWOOD f
l i
I I.
l Iiii
- i[1,L
!;Ifr; l
t!
'l'
>IJ{;
i,[
i>
.l.!i I!
(r f
,IIi*$
_.e a,
i D
1 O
O W
D I
_ e AR B
/
.g a
i f
.e 0
- i R
.s a
Y B
t re r
. o s
V
,e R
.t 1_
O c
P r
i e t.
f U it
.r H
'. t.t f[
T r
r
.r
.a L
a v
)
.e O
.t i.t 3
.u
.. e 8
(
B ei se q
a F
- e
[
s
,.t i.
i O
t
.e YT I
L is
..t
)
I e
d B
w.
3 e
A
. n.
J
,.'t yU G
u L
. e U
]I. t.t s
.i
=
n I
b1 N
.c.,
a i
A
(
e
.st
.,e
.t t t,
e t
V u
.-s
,t e
n A
v o
l e
t s
C U
.e
(
k_
8 O
R t
F r,
3 E
s r. 6 E
. e i.e i) c, g
R d..
p b t
i T
i.
r F
(
a L
r.,
t{
T
~
y.i.t ~
s,e A
gt rrt ie U
ae
.r n
F
~
si t
.i e
l i
W O
D T
w.
U l
S
_i a.__,
D L
.i
,u O
_e t
C r,
T A
r r
g j
l d. e.st
.t.
s,s)m R
_.e
- .c e
(
O l
r f.w.,
r.
- ,ei l
T C
w;.r.
I I_.t e. Y.w (-
g..
A
..i.
r.
E
.s
.c.
1 R
e
_i...., G) a,i I
e I
m,
_.st s
_in r
i J.1,l d n.
(,l i(
o i
~..
...)..
f..i..
t.(
.i l
_e.
e JGnc l4!ii!1!i!1 i1,,
ii
';i,l,
,ji Il
!li!iI1 l
l
l l
l O 1
1
..t..
6 : t a.
r t.,.. s. c :..
%. i t e s
/---
- r m
n
- O f
_I.
f-
--)
i&
i&
i&
E&..---
m 7.
,7
_,-.----9_-.-----~.
i (j
0 1 r - 1 -"- ~ "" - l- - - 7 i i g g 8 8 e I
i I
_J L.J L I
4 _ L 1_.
A_._L
, (/,6 [,
- p-D..
I
- o w
i
.: m,. a i
- 8satt'. val. C.s. vat.--**^r r t.., g. g a
3 l
e g.
L
~~WL
- E I
l
!
- g i i c t r s.
- m. f t se 1,,n,i,a
,..n,( ". )3
.t o# ut
(')
l c c.;,'."n t i c u i
i iu oc,tto r.
. (.. -.,. 7< < )
n t
i r<
a
.O i
i y
1 I
i eB_a_.
I i u s ou l I
e V
- it*'j
,f,
C,,1 t
.1.
WJL 5)
I.. s],.
.)
... s e.
-tl1
.s U.
.v r
s 5
5
.T
[E"l 5
'Yhh, g,
0
/2.
3 I
i r, *~I. r.i w i i r, u. vt
.: _ t_
c o, t>
.m.
"r t er r vaivr
..s o n c
- ,.<;.u..._
p,.o, 3.u,.
n o....... u c ou m,. mo
., i, o c 1.,,1 c,,
._;. -,... e.,-
c m -..
- s.,, i < v n s
.u g___.____',T~-~
.O p
.uic ictn.e e t.,
.uio, ci c. c I
d___d I
I
._I _.I
_ _ _ _ _-- I l
6 t
i l
l l
l.
l_l
' O i
L O
..,-i.1
.I.
l i
t s.
-r r, -
r n, e-
=
c..
Lc, O
.: t i.
.--1,
.mt u..a..- a 2-
. ;;.; a; i m. s. t
,.,.a.,
4 6T g
- 64b
%r;
.me a.
,, e-t.
. ~,,,..
O flg. 3.9 BYRO.'UBRAIC',-l00D - PORVs LOGIC DI AGPX O
O 37
.O an initiatir; hent for Cold cserpressurization; therefore, the impact of this failure is accounted for in the fault tree for both PORVs te open (Fig. 3.8).
3.3 Failure to Recover from Spurious Safety Injection
.O The failure to recover from a spurious safety injection appears in the event tree for Extended High Pressure Injection at Pa.ver (see Sec. 2.2 and Fig.
2.2).
As discussed in Section 2.2 this is an event that is neither extremely I4 r re nor difficult to control, and there are procedures for that recovery.
O Furthennore, since the operator has at least 20 minutes for recovery (as a
discussed in Section 2.2) this event is considered in this analysis as only a naderately high stress level event.
- O i
According to NUREG/CR-1278 the basic human error probability for this event is 0.003 and a multiplier of 2 is recommended for moderately high stress level.
In this analysis a value of 0.02 is used for the basic human error O
probability, and to be conservative a multiplier of 6 (instead of 2) was used.
16 Five people would be in the control room Three of the five are reactor operators (RO) and at least one of them has a senior reactor operator's (SRO)
!O license.
Tne remainina two are the shift engineers (SE, wno has an SRO license) and shift technical advisor (STA, who also has an SR0 license).
All of then, except for ow reactor operator would be totally involved with the f
cc et r ma erro b
t one use tne t rm as eo e d
O i
NUREG/CF.-1278 with the folloviing cepencencies among operators:
hign cependence (HD) between tne two reactor operators; moderate dependence (MD) l l
between the SE and tne first two; low deoencence between the STA and the fest.
- O Ine error f requency of the f our-ocrson team for this task (Recovery f rom Spurious Safety injection) would ce:
1 + 2.0E-2 1 +6 x 2.0E-2 1 + 19 x 2.0E-2 2.0E-2 x x
, ),)g_
j g 7
o
,O i
)
- O 38
_.-._-..,.,.,,.-m
,._.m,_
_-._m.,.
. - - _ _ _.. = = =.
i l
?O ibis value, i.lE 4, i t. used in the event tree for Extcr.ded High Pressure Ir.jection at Pc.ier at sht:an in Fig. 2.2.
O t
i l
i l g 10 I
i l
i I
l lO I
1 P
t O
e i
e O
t n
b i
O i
I i
i l
39 l
l www e..
.... w emme c,
=
l O
i A
O a
J The redian estimate for the frequency of liquid discharge from safety / relief valve in the Zion plant is 1.3 X 10-7 events / reactor year.
Table 4.1 breaks 4
fg down this result according to the three initiating events.
The frequency is l
dominated by contributions from the cold overpressurization tree event.
S/RV I
liquid discharge is thus an extremely unlikely event.
t 50 4.2 Byron and Braidwood Plants
)
,i i
The median estimate for the frequency of liquid discharge from a safety / relief l
valve is the Byron-Braidwood plant is 6.9 X 10-8 events / reactor-year.
Table
.O 4.2 breaks down this result according to the three initiating events.
The frequency is dominated by contributions from the cold overpressurization event.
S/P,V liquio discharge is thus an extremely unlikely event.
4 1
- O 4.2 General conclusions l
)
i The distnarge of liquid f rom safety relief valves in the Zion, Bryon, and Braidwood plants has been snown to be a possible but extremely unlikely event.
.O The estimated frequencies are basco upon conservative cata and assumptions and are suf ficiently low that even order-of-magnitude errors would not effect the j
cualitative conclusions.
O Furtner. tne scenarios of 5/RV liouid discnarge nave been predictec to occur iets frecuently tnan a small break LOCA, while tne consecuences (e.g.
nyoothetical overstressing of 5/RV discnarge piping) are certainly much less severe.
From tne point of view of safety risks, S/RV liouid discnarge appears to be an insignificant concern compared with LOCA events or FSAR transient
!!hile it is of cnurse aovantageous for S/RVs and associated piping tc events.
be verified operable f or a wice range of inlet conditions, it is eaually
}O imaartant tnat engineerina resources not be diverted from more realistic pursuits such as improvec 5/RV discnarge unoer expectec saturated steam inlet conditions.
O 40
i 4
jGD T AE,LE 4.1 l
SUMMARY
OF RESULTS FOR ZIch PLAST FREQLELOY i
i i
0F LIQUID DISCHAP.GE FEOM SAFETY / RELIEF VALVES
- Gl 1
Calculated Frequency Of Occurrence Type of l
(Events / Reactor Year)
Discharge Initiating Event t
3.6 x 10-8 Steam followed by l
!I' Extended High Pressure saturated or slightly i
l Injection subcooled liquid; possible valve cycling I
4 9.6 x 10-b Far subcooled liquid l
Cold Overpressurization g,
5.4 x 10-10 Steam followed by Main feedsater Pipe saturated liquid Rupture i
jet i
i i
i f
1.3 x 10-7 f
lgp TOTAL i
l e
O i
I e
l L
e 1
O 41
i l
i l
I I
i TABLE 4.2 l
l
' O SU!?'ARY OF RESULTS FOR BYRO ;/SRAID'.!00D PLA!!IS i
FREQUEfiCY OF LIQUID DISCHARGE FROM SAFETY / RELIEF VALVES l
I Calculated frequency
.g Of Occurrence Type of j
i Initiating Event (Events /ReactorYear)
Discharge Extended High Pressure 9.9 x 10-9 Steam followed by i
Injection saturated or slightly i
subcooled liquid; possible velve cycling j
Cold Overpressurization 5.8 x 10-8 Far subcooled liquid i
5.6 x 10-10 Steam followed by fiain Feedwter Pipe l9 Ruptu re saturated liquid i
O l
l l
e.9 x 10-8 TOTA:.
l I
i I
\\
O I
l l
I 2
i l*
i t
i
)
i is i
42
/
/
9 REFU:ECES
/
n..
~
f;.
,I 1.
NUREG-0737, "C;arification of TMI Attien Plan Ret,uirements,"
3 n',-
O hem 11.id.1^, Noven.ber 1980.
2.
W.R. Hr4 Ling,.et. al., "EPRI/CE PWR Safety,Vgive Test Program,'
. 3/
y EPRI Research Project V102-2,. Fin 31 Report, July 1982.
d (-
D. F. Streint,,5."EPRl/CE PWR Safety and id, D.6f' Valve Test Program, Uptream 3.
O,
~
Pressure O';ctllations," Combustion Erdin*ering; Lettgr PE-81-415, December
,' a s
p-n;,
18, 1981.
?
9c
.,,. y
._ u 4.
B. R. Strong, 'et.a'.<,yteam Har:ner Design Loads for Satety@elief Valve
~
Discharge Piping," 'f rom kf et'y Relief Valves, ASME, 1979.-
O B.S.1 jnger and R. S. Mayl "/palysis of Safety /Reliev,hlve,Chatteh a
3 5.
and Transient Problems," sal' jport to A'nmonwealth Edison, JuneL isB2.
-1 D. Harris, et:al, "Proba'aflistic {valtration of High Pressure >Llquidl 6.
.m O<
ChaMense of Safety l/ Relief V.alve Piping," SAI. report S6!;245-81-PA submitted to BWR Oyner's-group a.nd General ' Electric Coinpany, April la31, _
m, 7.
A. Meliksetian and,A'KJ. Sulencar,"."JEive~ Inlet Fluid'Coeditions for; Pressurized Safet? and_ Relief Valv;; in Westingh0Vsa-Designed Flants",
A l
- O EPRI research prdieft'Y102-19,sJanuary 29~ 1932.
./
e i
8.
Zion Probabilistic SafetL Study.
CommonwealthEdist(Company.
M l
~
9.
Reactor Safety Study:
An Assessment of Accident RfDs in U. S.j; Commercial Nuclear Power Plants, "NUREG-75/rii4, Gi toodr 1975.
s O
10.
A".'- 0. Swain and H. E. GgtmarI, "fiaridbook c4,M*m i6.ii i abf l i ty -
7-
-u,
!Analnis ~with Emphasis 6r Nuclear'Pcwer f lant AhDLicatiohs,'l Y
r 2
r.
GURES/CR;127s, Octocer_193C.
Hubrie and C. F. Miller, " Data Summartes of Licensee tvenL ReDdE 11.
W. E x
~
0 of volves at U.S. Comerclei huciear Pc.;e-Plants," NUREG/CR-1363;' June i
l 19E,. >
12.
Zion' Station, Units 1 and 2, Final Safety Analysis Report.
[
Cod.touweal th Edi son - Companyw Decembe r 197C.
l I
O 13.
Byron /braicwoodFina'lSiMtfAnalysesRecor..
Corrnannealin Edison
^
'uc: pan)L,,, _
1,.
Zion t.mwye Jcy and Goeratioits Pi;ccecure..
9
-Saf etgj nje ninc/ Accicdct DiF;nostic:
EOF-0:
[
a G05-1:
Piant Start 0r.
"G00-2:
Fiant'Snutco,,r
/
.N, 4
- ~/
~
43
~
--ye-w
-er e,
e--w---.he---,,-
-,r-.-,-w-r r
_..~_- _.-_____ - - - -
t I
- 9 l
l l
- 15.
E. W. Hagen, "Co.mpressed Air and EscLup in I;uclear Paier Plants," F.epart I
d.g!v' by OR ;'. on a contract for f,RC, to be published.
t e
16.
Private Ccrunications with G. Klopp and F. Highland (Com.cnwealth Edison lg7 Company).
i h
4
{/
- e t/
t -L.
j9 l
I I
!9 u
J' A %
h
- 9 I,
w v
- )O
'I i
(
I L
b le t
I 9
I a
l 4
IO l '.
i i
t i
4 IG f
1 4
9 0 44
.