ML20206K752
| ML20206K752 | |
| Person / Time | |
|---|---|
| Issue date: | 09/09/1988 |
| From: | Beckjord E NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Stello V NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| References | |
| REF-GTECI-125, REF-GTECI-NI, TASK-125, TASK-OR NUDOCS 8811290524 | |
| Download: ML20206K752 (2) | |
Text
.
e.
Q 0 0 6, Oek d c1 2M. T SEP 9 198g i
I MEMORANDUM FOR:
Victor Ste'lo, Jr.
Execut'7e Director for Operations i
FROM:
Eric S. Beckjord, Director Office of Nuclear Regulatory Research L
SUBJECT:
RESOLUTION OF GENERIC ISSUE 125.11.7, "REEVALUATE PROVISION TO AUTOMATICALLY ISOLATE FEEDWATER FROM STEAM GENERATOR DURING A LINE BREAK" 7
The subject generic issue has been resolved and requires no further study.
It relates to the concern that the automatic isolation of Auxiliary feedwater (AFW) design feature used in twenty-seven PWR's might have a sicjnificant negative net impact on safety by compromising the AFW function with inadvertent actuation of automatic isolation.
This issue was identified as a potential l
generic issue after issuance of the Augmented Inspection Team report on the t
Davis-Besse event of June 9, 1985.
It was subsequently evaluated and assigned a HIGH priority classification.
The enclosed regulatory analysts of the AFW autornatic isolation feature, NUREG-1332, which is partially based on the technical findings reported by Itlaho National Engineering Laboratory in NUREG/CR-5178 (Enclosed), shows that for the postulated removal of the AFW automatic isolation feature in the plants ar,alyzed, (a order of 10") core damage events per reactor year and (b) the risk reductio the reduction in core damage frequency (C0F) would be in the would be about 40 person-rem per plant.
Furthermore, for some plants, it is expected that removal of the autornatic isolation of the AFW system would result in an increase in risk.
This risk increase is particularly applicable to plants with no-flow restrictors in the AFW pump discharge lines.
Bastd on the enclosed regulatory analysis and its supporting documentation, we conclude that removal of the AFW automatic isola Hon feature will neither i
resu;t in a substantial safety improvement, nor will it be cost-effective.
Hence, Alternative Resolution No.1 "No Action," as recorarwnded in the encicsed regulatory analysis, is adopted as the appropriate resolution of this generic issue in accordance with the Backfit Rule, 10 CFR Part 50.109(a)(3).
Consistent with the Starviard Review Plan (SRP), the "No Action" alternative does t'ot precluae a licensee from proposing to the NRC staff the removal of the AFW automatic isolation feature based on plant-specific considerations.
Resolution of this generic issue has been accomplished in accordanco with the proceduces contained in RES Office Letter No. 3, Procedure and Guidance for the Resolution of Generic Issues.
(f*)f es112905c4 000909
' \\
PDR QTEC1 GNIl
GEP 9 1988 V. Stello liUREG/CR-5178 has been issued to provide a record of technical findings on the safety significance of the subject generic issue, which is considered resolved.
It will be removed from the list of active generic inues, and the section of fiUREG-0933 dealing with GI-125.II.7 will be revised ar.cordingly.
\\s\\
Eric S. Beckjord, Director Office of huclear Regulatory Research
Enclosures:
1.
Regulatory Analysis of GI-125.II.7, NUREG-1332 2.
ilVREG/CR-5178 cc:
E. Jordan, CRGR R. Fraley, ACRS F. 0...espie, NRR B. fiorris, RES C. Sheron, RES G. A, lotto, RES DISTRIBUTION:
RES Circ RES Chron DCS f(L/007 DSIR C/F RPSIB R/F D. Basdekas G. liazetit,
- k. Kniel W. Itinners R. W. Ifouston T. Speis E.Beckjord PDR S
PSIB DSIR:
IB QSJ 4L 0 }bD OFC:
DSilJ1 NAliE :
DBa jf etis XXniel Wl11aners
,R Ouston DATE:
9/2/88 (j/p/88
<//Z/88
([ /(, /88 p'/ f/88 h
OfC GRID RES:D flAliE:
TSFe,s>
EBec kjord DATE:
Q 88 C////88 "0FFICIAL RECORD COPY"
ItVREG-1332 REGULATORY ANAL).*S FOR THE RESOLUTI0li 0F GENERIC ISSUE 125.11.7, REEVALUf 'E PROVISION TO AUTOMATICALLY ISOLATE FEE 0 WATER FROM STEAlt GEtiERATOR DURING A LIllE BREAK SEPTEMBER, 1988 O. S. NUCLEAR REGULATORY C0!Vi!SS10ll 0FFICE OF fiUCLEAR REGULATORY RESEARCH PREPARED BY:
DEMETRIOS L. BASDEKAS, SR. TASK MAfiAGER h
ABSTRACT Generic Issue 125.II.7 addresses the concern related to the automatic isolation of auxiliary feedwater (AFW) to e steam generator with a broken steam or feedwater line.
This regulatory analysis provides a quantitative assessment of the costs and benefits associated with the removal of the AFW automatic isolation and concludes that no new regulatory requirements are warranted.
I i
t iii
,9..
_<w
7
(
L I
TABLE OF CONTENTS Page L
L ABSTRACT............................................................. 111 j
LIST OF TABLES........................................................v11
[
PREFACE...............................................................ix EXECUTIVE
SUMMARY
.....................................................x!
- 1. STATEMENT OF PR0BLEN..............................................
1 l
- 2. 0BJECTIVE..........................................................?
- 3. ALTERNATIVE RESOLUTIONS...........................................
2
- 4. T ECHNICAL F IND INGS S UMMARY........................................ 2 4.1 S u mma ry o f R i s k E s t i ma t e s.................................. 2 4.2 Summary. of Implementatien of Alternative Resolutions....... 5
- 5. CONSEQUENCES......................................................
5 5.1 Alternative Resolution No. 1 - No Action................... 6 5.2 Alternative Resolution No. 2 - Removal of Automatic I s o l a t i o n o f A FW Sy s t e m.................................. 7 5.2.1 Change in Ri s k Eva lua tion........................... 8 5.2.2 Cost Estimation.....................................
9 5.2.3 C o s t - B e n e f i t S u mma ry................................ 9 5.2.4 Cost-Benefi t Uncerta inties.......................... 11
- 6. DECISION RATIONALE................................................
15 7.
IMPLEMENTATION....................................................
15 REFERENCES...........................................................
16 V
9 LIST OF TABLES Page Table 1.
Change to CDF Caused by Removing the AFW Automatic I s o l a t i o n Sy s t e m......................................... 4 Table 2.
Discounted Present Value of Averted Onsite Damage l
Cost per Plant.....................
..................... 7 Table 3.
Risk Change Due to Proposed AFW System Ho d i f i c a t i o n............................................. 9 Table 4.
Cost Estimate for Proposed Hodifications
( p e r P l a n t ).............................................. 10 r
Table 5.
Summary of Cost-Benefits in Dollars per i
Person-rem Reduction.....................................
11 Table 6.
Basic Data Used in the Uncertainty Sensitivity A n a l y s i s................................................. 13 Table 7.
Sensi tiv i ty Ana ly s i s Resu l ts............................. 14 I
t i
vii
PREFACE This report presents the regulatory analysis, inc;uding the cecision rationale, for the resulution of Generic Issue 125.11.7.
The objective of this regulatory analysil is to determine whether the removal of the auxiliary feedwater (AFW) automat'c isolation feature, as proposed as a result of the review of the Davis-Betse event of June 1985, it warranted. As part of this effort, four plant deLigns were evaluated.
The risk change est.imates, cost-benefit analyses, and other insights Sained during this effort have shown that no new regulatory requirements are warranted in rela:1on to this generic issue.
Deraetrios L. Basdekas ix
EXECUTIVE SUlWARY The report provides supporting information, including a cost-benefit analysis for the fluclear Regulatory Commission's (NRC) resolution of Generic Issue 125.11.7, "Reevaluats Provision to Automatically Isolate Feedwater from Steam Generator During a Line Break." This issue addresses the concern related to the automatic isolation :f AFW to a steam generator with a broken steam or feedwater line. Following the Davis-Besse event in June 1985, the benefits of this automatic isolation oesign feature system versus its disadvantages were questioned. The purpose of this generic issue is to reevaluate this design feature and to determine whether its disadvantages may outweigh its benefits.
The benefits of automatic AFW isolation in the event of a steam or feedwater line break include:
1.
The steam generator inventory blowdown is minimized. While isolating AFW does not prevent the initial secondary side inventory blowoown, it does prevent continued blowdown af ter the initial inventory is expended, ano thus minimizes the use of condensate supplies.
2.
Excessive primary system cooldown is minimized, thus reducing the likelihood of recriticality and pressurized thermal shock.
3.
Excessive containment pressure for a steam line break inside containment is minimized by discontinuing feedwater adoition.
The disadvantages of automatic AFW isolation are related to a concern that the automatic isolation system may reduce the reliability of the AFU function.
Failures that cause inadvertent AFW isolation could cause loss of all AFW system flow during accidents or transients. Additionally, during a controlled r
cooldown af ter an accident, the thresholds for autecatic AFW isolation (such as low steam generator pressure, or high steam generator to steam generator pressure differential) may be crossed, which would require that the operator lock out the isolation logic as the steam generator parameters approach the isolation setpoint.
Failure to do so would defeat the AFW function until the operator intervenes.
There were four plant designs evaluated under this effort.
The results of the quantitative estimates of risk changes in these four plant designs show that removal cf the AFW automatic isolation feature would result in a small risk increase in two plant designs, while It would result in a small risk (CDF)reductionwouldbeintheorderof10'yercasesthecoredamagefrequency reduction in the other two.
For the two lat
/Rx-year and the risk reduction woulo be ahott 40 person-rem per plant.
The cost-benefit ratios for f.taking necessary modifications range from 58,000/ person-rem to $17,000/persun-rem depenoing on whether the 27 plants affected by this issue already have flow restrictors in the AFW pump discharge linet, j
x1
~
s Based on the results of the cost..oenefit analyses, and the insights gained during assessment of the pros 6nd cors of removing the automatic isolation design feature of the AFW system, we conclude tha*. no backfit requirement to remove this design feature from PWRs is warranted in accordance with the Backfit Rule, 10 CFR Part 50.109(a)(3).
1 i
.1 4
xi1
REGULATORY ANALYSIS FOR THE RESOLUTION OF GENERIC ISSUE 125.11.7, REEVALUATE PROVISION TO AUTOMATICALLY ISOLATE FEEDWATER FROM STEAft GENFRATOR DURING A LINE BREAK 1.
STATEMENT OF PROBLEM The NRC requirements for the auxiliary feedwater (AFW) system related to this issue, as stated in the Standard Review Plan (SRP) (Ref. 1) provide that "the system design possesses the capability to automatically terminate auxiliary feedwater flow to a depressurized steau generator, and to automatically provide feedwater to the intact steam generator. Or, as an
.:ternative, if it is shown that the intact steam generator will receive the minimum required flow without isolation of the depressurized steam generator and containuent design pressure is not exceeded, then operator action uay be relied upon to isolate the depressurized steam generator." Based on these requirements, 27 PWks have an AFW automatic isolation design featur? to assist in mitigating a steam or feedwater line break.
Generic Issue 125.11.7 addresses the concern related to the automatic isolation of AFW to a steam generator with a broken steam or feedwater line.
Following the Davis-Besse event in June 1985 (Ref. 2), the benefits of this autonatic isolation system versus its disadvantages were questioned (Ref. 3).
The purpose of this generic issue is to reevaluate this design feature and to determine whether its disadvantages may outweigh its benefits.
The benefits of rrcomatic ArW isolation in the event of a steam or feeowater line break are:
1.
The steam generator inventory blowdown is minimized. While isolating AFW i
does not prevent the initial secondary side inventory blowdown, it does l
prevent continued blowdown after the initial inventory is expended and thus minimizes the use of condensate supplies.
2.
Excessive primary system cooldown is minimized.
Continued AFW flow to the steam generator with the break would result in a sustained cooldown and subsequent repressurization of the primary system, thus contributing to pressurized thermal shock conditions.
As the primary system cools down l
because of the heat transfer to the depressurizing steam generator, a reactor recriticality could occur, especially if the reactor core is approaching end-of-lite. This would introduce thermal energy to the primary system. Shutting off AFW to the faulted steam generator will reduce the likelihood of recriticality.
3.
Excessive containment pressure is minimized.
The containment is designed to accommodate the pressure increase caused by a primary system loss-of-coolant occident (LOCA). A steau or feedwater line break within containment might cause the cc.itainment design pressure to be exceeded if the automatic AFW isolation were not available, l
1 k
j 4
In some plants, the AFW automatic isolation feature is required to divert j
AFW from the affected steam generator for orderly and safe plant shutdown and to meet the single failure criterion in supplying feedwater to the intact steam generatorts).
l The disadvantages of automatic AFW isolation are related to a concern that the automatic isolation system may reduce the reliability of the AFW function.
Failures that cause inadvertent AFW isolation could cause loss of all AFW system flow during accidents sr transients. Additionally, during a controlled cooldown after an accident, the thresholds for automatic AFW isolation (such as low steau generator pressure or high steam generator to steam generator pressure differential) may be crossed, which would require that the operator lock out the isolation logic as the steam generator parameters approdch the isolation setpoint.
During such a scenario (not requiring isolation), the accompanying distractions could result in a failure to lock out the automatic isolation.
Thus AFW would not be available until the operator ir:tervenes.
The focus of this ger.cric issue is to reevaluate the automatic isolation of AFW by addressing the question as to whether the positive aspects of this design feature are outweighed by the negative aspects.
4 i
2.
OBJECTIVE The objective of this Regulatory Analysis is to determine whether the proposed change in the AFW automatic isolation design, as proposed in the af termath of the Davis-Besse event of June 1985 (Refs. 2 and 3) is warranted.
3.
ALTERNATIVE RESOLUTIONS There were two basic alternatives (Ref. 3) considered as a basis for resolution of Generic Issue 125.11.7.
l 1.
Take no action. Consistent with the SRP, this alternative does not preclude a licensee from proposing to the NRC staf1 the removal of the AFW automatic isolation feature based on plant-specific considerations, or 2.
Remove the automatic isolation feature of the AFW system.
Alternative No. 2 would entati the removal of the automatic isolation feature with the attendant reliance on operator action to cope with a steam or feedwater 4
line break, and, for plants that do not have them already, the
[
installation of flow restrictors in the discharge lines of the AFW pumps for pump runout protection.
4 TECHNICAL FINDINGS SUMARY, 4.1 Summary of Risk Estir.16tes i
Probabilistic tre nods were used for a generic study of the changes in risk I
that would occur if the AFW automatic isolation feature was removed.
The results of this study by the Idaho Nstional Eng' neering Laboratory (INEL) are i
repcrted in NUREG/CR-6178 (Met. O.
1 Four PWRs, one. from eacn reactor vendor (two BaW AFW designs), were evaluated to determine the AFW automatic isolation system's contribution to the core damage frequency (C0F).
Since the decay neat removal safety finction via cooling through the steam generator is of importance in *.hese COF s. ouences, it is expected that the greatest risk associated with this issue would ' e for o
plants with marginal or no feed-and-bleed capabilities. An analys4 has been performed to determine the sensitivity of CDF changes, for the sequences analyzed for this issue, to the unavailability of the feed-and-bleed capat,i li ty.
The results of this sensitivity analysis confirm this conclusion (Ref. 5).
This study included two such plants.
Another significant consideration was that some plants utilize the automatic AFW isolation system to prevent AFW pump runout conditions with resultant possible pump damage and AFW system failure when supplying water to a depressurized (steam or feedwater line break condition) steam generator. The evaluation performed for this study clso included three of these plants.
The evaluation indicates that the quanthative effects on CDF of the AFW automatic isolatior ;ystem are largely dependent on the particular plant design.
The auxiliary feedwater systen is a part of the balance of plant cesign usually contributed by the architect-engineer, and not normally part of the more standardized reactor and primary system provided by the reactor vendor. The estimateo changes in CDF due to AFW isolat;on system's contribution are low, but t;1e range between the highest and lowest value is about an order of magnitude.
At Plant A (CE), removing the isolation system will not cause a failure of 4
the AFW system because the pla,1t has flow restrictors ir the AFW headers that limit flow to a ruptured steam generator and maintain flow to the intact stesm generator. Removing the isolation system at this plant would decrease the CDF oy 5.5E-07. At Plant B (B&W), removing the automatic AFW isolation system would cause AFW system failure without operator action bt.cause the plant does not have flow restrictors in the M headers and the pump trains are cross connected.
Thus, AFW flow would be liverted to the ruptured steam generator and the pumps would be damaged because of low net positive suction head caused by the high flow rate.
Removing the automatic isolation system woulo cause a J
C0F increase af 9.6E-08. A: elant BB (B&W), removing the automatic isolation system would also cause AF'/ failure without operator action because the plant does not have flow restrictors in the AFW headers and the pumps are cross-connected.
Removing the automatic AFW isolation system would cause a C0F i
l decrease of 9.0E-07. At Plant C (W), only part uf the AFW system would fail if the automatic isolatior, features were removed and the operator took no action to isolate a ruptured steam generator.
A ruptured steam generator would cause the flow from one of the t.otor-driven pumps ano the turbine-driven purup to be diverted to the break, but the other notoredriven pun) would still supply two intact steam generators, which is the AFW success critsrion for rnost accidents.
Removing the automatic isolation systen would cause a CCF increase of 4.0E-07.
3 i
i
Table 1.
Change to CDF Caused by Removing the AFW Automatic isolation System Decrease in C0F Increase in CDF C:ured by Deactivating Caused by Deactivating Het Chenge Plant AFW Isolation System AFW ! solation System In C0F A (CE) 5.50E-07 0
-5.bE-07 8 (B&W) 4.40E-08 1.4E-07
+9.6E-08 BB (B&W) 1.04E-06 1.4E-07
-9.0E-07 0 (W) 4.00E-08 4.4E-07
+4.0E-07 Even though only four plant designs were evaluated in tnis study, it is expected that the results of this study can be extrapola'.ed to address this issue on a generic basis.
This judgn.ent is t; ed on the following factors:
1.
The four pl6nts evaluated in this study include plants with somewhat different automatic AFW isolation system designs.
Based on a survey of til PWRs, one of these designs is representative of system designs used at all the 27 plants that have the AFW automatic isolation feature.
2.
If automatic AFW isolation designs are used at some plants that may be significantly different from those evaluated in this study, the findings of this study related to differences in CDF compared to systeu design can be extrapolated.
This study indicated that the differences in isolation system design had little bearing on the change in C0F.
The major factor affecting the C0F calculations was the presence, or absence, of flow restrictors in the AFW system. All PWRs either have AFW flow restrictors or de not.
This study showed the worst case was for plants that do not have flow restrictors.
Even these plants showed no significant benefit in removing the AFW automatic isolation feature.
i 3.
As noted in Item 2 above, removal of the AFW automatic isolation system would result in thi largest risk increase for plants that do not have AFW flow restrictors.
lf the existing isolation system were renoved, these i
plants would incur the highest cost because AFW flow restrictors would need to be added.
However, the most favorable cost-benefit ratio i
(approximately $8K/ person-rem) calculated during this study used the highest risk reduction value calculated for a plant without AFW flow restrictc:
and used the least expensive cost (for plants not requiring i
the addition of AFW flow restrictors).
This conservative methoc was used l
for initial scoping calculations to account for analytical uncertainty ard also to provide some assurance that offferences not addressed in Items 1 and 2 above would not change the conclusions of this study.
t
[
i
e 4.2 Summary of Implementation of Alternative Resolu tions Based on the results sumarized in Subsection 4.1, the implementation of the two alternatives discussed in Section 3 may be summarized as follows:
1.
The "no-action" alternative would simply rely on normal distribution of this regulatory analysis and the contractor's report (fiUREG/CR-5178) to allow these insights to be available to all utilities.
ho further action would be taken.
Consistent with the SRP, a licensee may propose to the NRC staff the removal of the AFW automatic isolation fe.iture based on plant-specific considerations.
2.
The c. <. alternative would entail the removal of the automatic isolation feature with the ottendant reliance on operator action to cope with a steam or feed line break, and, f ar plants that do not have them already, the installation of flow restrictors in the discharge lines of the AFW pumps to protect them frcm runout.
5.
CONSEQUENCES This section assesses the cost-benefit aspects of the 6lternative resolutions of this issue.
In such an assessment, "costs" provide a measure of primarily economic consequences resulting from the implementation of alternative re elutions.
Based on their definition, these costs may be considered positive le.g., the incurred costs in 1988 U.S. dollars for installing, operating, and maintaining any replacement power during a necessary downtime for the plant (s)g the co the plant modifications needed to implement a resolution, includin), or they may be considered negative (e.g., savings to the operating utility in terms of averted accident costs associated with plant repairs, cleanup, power replacement,etc.).
Thus, the net cost represcnts the positive costs minus the present worth of the negative costs (averted ensite costs) over tha remaining lifetime of the plant (s).
Conversely, the term "benefits" denotes the improvements made to public health and safety as measured in the reduction of person-reas of population exposure os well as in the offsite property damage costs associated with land interdiction and decontamination that may be necessary.
The number and importance of parameters contributing to the costs and benefits vary with the type of accident and the plant location (Ref. 6).
In the analyses perfonted in this study, the site characteristics for a "typical" midwestern plant and site (Rets. 4 and 7) are usert because they are representative ut the population of U.S. PWR$ affected by this issue.
This consequence analysis contains Joue bounoing assumptions on the consequences associated with individual plants and the corresponding analyses on a plant-specific basis. Howeer, the neeu to go back and perfurm a more realistic cutt-benefit estimata became moet when, even with the preceding bias, the results did not warrant a bockfit.
5
The averted onsite dose due to avoided accidents associateo with the removal of the AFW automatic isolation feature (for plants that show a risk reducyion) is neglected in this analysis.
This is because the change in CDF
( 10- events /Rx-year) and corresponding total consequences are small, thus the
- s. all onsite population and attendant meteorological characteristics would make the onsite averted dose negligible.
Also, based on the small ulculdted changes of CDF, the averted onsite costs involved would be very small (ee Section 5.1) compared with the costs dssociated with the removal of the automatic isolation capability.
- Hence, these small costs were also neglected in the calculation of the respective cost-benefit ratios.
The expected uncertainties associated with offsite consequences regarding perron-rem cases with offsite protective actions are large (Ref. 8).
Therefore, for initial scoping of cost-benefit estimates, the analysis was based on a 50-mile raoius of a "typical" midwestern location (Refs. 4 and 7).
As it is shown in Section 5.2.3, the need to obtain a more re611stic cost-benefit estimate became moot when, even with the preceding l
beunding bias, the results dir not warrant a backfit.
5.1 Alternative Resolution No. 1 - No Action There are normally no costs attributed to a "no-action" alternative.
However, by cvnvention they are treated as benefits of other alternative resolutions. Averted onsite costs can have a significant effect on the overall cost-benefit ratio depending on the expectea reduction in C0F.
These costs may be estimated by multiplying the change in accicent frequrney by the discounted l
onsite damage costs. The following equations (Ref.10) may be ured to estimate theie costc or. a per plant basis:
j V
= N a FU gp and U = C ((e #t )/r ][1 - e #It ~t )](1 - e #5) i f
i m
i where P = value of onsite averted custs V
=
number of affected facilities = 1 N
LF = reduction in accident frequency = 10~7 U
= present value of cnsite property damage l
C
= cleanup, repair, and replacement power costs = $2.4 billion t
= years remaining until end of plant life = 25 i
g
= years before reactor beg) ins speration = 0 t g
= discount rate =.10 (10% ;.05 (5%)
r
= period of time over which damage costs are paid out (recovery m
period in years) = 10 On the basis of the above fornuias and assigned paraneter values, we find that the averted cnsite costs per plant are expected to very between $1,500 and 53,000.
Table 2 sun.,arizes the discoui.ted voiues foi' 101 and 51 discoant rates.
6 I
i
Table 2.
Discounted Present Value of Averted Onsite Damage Cost per Plant Discounted Present Value Averted Costs 10% Discount Rate 5% Discount Rate Cleanup, repair, and
$1,500 53,000 replacement power As shown above, these averted onsite cotv.s for this particular issue are very small ($3,000 or less).
In some cases, they are negative. As discussed in Section 6, these results t.ranslate co retaining the curreat design as the best alternative.
5 5.2 Alternative Resolution No. 2 - Removal of Automatic Isolation of AFW System This alternative resolution involves the removal of the automatic isolat'on of AFW during a steam or feed line break.
A detailed analysis of representative PWR designs was performed by selecting one Westinghouse plant, one Combustion Engineering plant, and two Babcock & Wilcox design versions of the same plant (with and without the Emergency Feedwater Initiation and Control (EFIC) System).
As propored originally (Ref. 3), removing the AFW automatic isolation feature consists of disconnecting and securing the automatic enable circuits.
This modification will provide the AFW system with manual control once the 2
l system has been activated. Once the isolation function becor s a manual action, it will necessitate additional operator training anc revised plant 8
operating procedures. A further complicction has been identified for plants that use the automatic AFW isolation system to prevent AFW pump runout.
If the autornatic AFW isolation system were disabled in these plants, additional plant modification would be required to prevent pump rN 9t.
A survey performed W.11 operating pWRs indicated that 27 plants woulo bo affected by this issue. Nineteen of these 27 plants would be affected by AFW pump runout considerations. These plants, if they have the AFW automatic 1
isolation removed, would require the additional installation of AFW pump discharge flow restrictors or block existing suitable valves to prevent pump
)
runout.
Changes to the AFW flowrate characteristics and initiation times by the
)
introduction of flow restricting cevices and operator action, respectively, would also require detTiled reanalysis of steam and feed line break accidents
)
in the Final Safety Analysis Reports (FSARs), changes of operating procecures, and operator training. Technical Specification changes would be required to reflect the modified design and to provide for periodic testing of the modified j
AFW systen.
I
)
i 1
5.2.1 Chance in Risk Evoluation To evaluate the proposeo modifications on a risk change (benefit) versus cost basis, the risk charge associated with the scenarios of concern was calculated, using the fol. lowing relationship (Ref. 4):
Change in CDF (events /yr) x Containment Failure Probability x Offsite Radiation Dose (person-rem) = Risk Change (person-rem / year)
To calculate the total change to the potential population exposure or risk per plant life due to this issue, the above relationship was extended over the plant life, taking into account plant downtime.
The total change in population exposure over the remaining plant lifetime is calculated as follows:
Change in Risk (person-rem / year) x Remaining Piant Life (years) x Plant Utilization Factor = Total Change in Population Risk (person-rem)
The potential change in risk due to the propossd AFW modification, for the selected plants, was evaluated using the plant-specific containment failure and release category information delineated in the respective PRAs (Ref. 4).
To extrapolate the estimated person-rem / year risk to total change in plant risk, the plant life was estimated utilizing the expected remaining lifetime of 23 years. An associated utilization factor of 75% was used.
These values were taken from 14UREG-0933 (Ref. 7) and Reference 11.
The plant-specific PRAs (Ref. 4), except for plant C (W), were performed on the basis of the methcdology and data developed as part of the WASH-1400 analysis, which has been subsequently found to be conservative in pcrtions of its estimation of the source terms and containment failures.
However, ongoing work has also shown areas in which the WASH-1400 treatment was not sufficiently conversative; for example, neglecting to account for the effect of accident (Ref.12) plated out fission produ ts in the initial stages of an revaporization of Hence, the conservative elements tended to be balanced out by other elements that were on the nonconservative side.
The plant C (W)
PRA (Ref.12) was perfors'.ed on the basis of the methodology and cata developed as part of the Work on filREG-1150 (Ref.13), i.e., the applicable release fractions were taken from the 19 release bins instead of the nine release I
categories used in WASH-1400 and the PRAs for Plants A, B, and BB.
The final numbers for the change in risk for the four plant designs analyzed are within a factor of about 3 (Ref 4), which is well within the error band of of 210 that was used in the cost-benefit sensitivity analysis.
Hence, while extensive work is under way to better quantity the source termr. of various accidents, the WASH-1400 results in this area for Plants A, B, and BB were deemed appropriate.
Table 3 summarizes the change in risk due to the AFW system modification, 8
Table 3.
Risk Change Oue to Proposed AFW System Modification flVREG-0933(7)
Plant PRA Data Data Data Plant (person-ren)
(person-rem)
A(CE) 36(Decrease 1.5(Decrease)
B (B&W) 4 Increase 0.25(Increase)
BB (B&W) 44 Decrease 3.0 (Decrease)
C[W) 13 Increase)
No Data 5.2.2 Cost Estimation using the guidance contained in References 10 and 14, the costs of implementing the modifications involved uncer Alternative Resolution No. 2 *:ere analyzed in NUREG/CR-5178 (Refs. 4 ano 9). A cost analysis for disabling the automatic feedwater isolation system was also conducted by the NRC staff, as documented in a memorandum from A. J. Dipalo to G. R. Hazetis, dated February 5, 1988 (Ref. 9).
The results of these two cost analyses were in close agreement. Table 4 presents the results of the cost analysis (Refs. 4 and 9) with the exception that replacement power costs were also calculated for the case of installeo flow restrictors.
5.2.3 Cost-Benefit Surnary To determine the cost effectiveness of the proposed modification for each of the plants, a cost-benefit analysis was performed. The cost-benefit ratio l
l was estimatt.d tsing the following equation:
[EstimatedCostofModification(S)]i[Changein l
Risk (person-rem)]=[ Cost-Benefit ($/ person-rem)]
The risk reduction values employed in this analysis used the largest decrease in risk from Table 3 and the smallest cost from Table 4.
This scoping approach was taken to ensure that the likelihood of excluding viable fixes was minimized. Should the result be shown to be cost-beneficial, a new estimate would then be necessary to obtain a more.ealistic value.
The results of this analysis are shown in Table 5.
The cost-benefit analysis was compared against the $1,000 per person-reu screening value to evaluate the cost effectivenets of the proposed modification. As snowr neither set of modifications would be cost-beneficial.
9 l
[
Table 4.
Cost Estimate for Prcposed Hodificatioris (per Plant)#
~
Cost to Disable Automatic AFW Cost to Disable Automatic AFW Isolation Systems Without Flow Isolation Systems with Flow Cost Category R_estrictors Installed ($1000)
Restrictors Installed ($1000) b Design, Hordware, and Installation Hot Applicable
$75 C
d d
Utility Licensing
$250 800 Operotor Training 43 43 t4RC Review 58 158 Totol it modifications are performed
$351
$1,076 during a scheduled outoge Replacec.esit Power Cost Not Applicable 16,000" Total if modifications are performed
$351
$7,076 d.
Without overted onsite costs (see Section 5.1).
b.
Estisaote includes design, installation, calibration, and testing.
c.
Estination includes Technical Specification, FSAR, and procedure changes ond amendments.
d.
Estinate nosed on reanalyses required of selected DBAs.
Estinate bcsed on the power replacement costs of $300K/ day i
e.
dssociated with a 20-day specially required outage (Ref.14).
Table 5.
Sunnary of Cost-Benefits in Dollars per Person-rem Reduction for Plants Not for Plants Requiring Requiring Hardware Hardware Modifications Noaifications
$351K/44.4 5768K/44.4 person-rem =
person-rem =
57,905/ person-rem
$17,290/ person-rem l
5.2.4 Cy.-Benefit Uncertainties to some uncertainty.y used to calculate the cost-benefit ratios are subject The cost estima These costs were estimated using NUREG/CR-4568 (Ref. 14) as guidance and the specific information provided in Reference 9.
One major area of uncertainty is whether the proposed modifications can be completed during a scheduled outage. Table 4 shows the costs associated with the bounding cases (i.e., estimated cost when the modification requires an outage--the upper bour.d cr. esti=ted cost, and the estimated cost when the modification is perfortred during a scheduled plant outage--the lower bound on estimatedcost).
exclusion of any avertea onsite costs (see Section 5.1) g those related to the Other uncertainties, of lesser in:portance, includinare not large enough to cause any change in the conclusion regarding the cost effectiveness of removing the AFW automatic isolation feature.
The individual tasks perfurmed during the technical evaluation of GI 125.11.7 are also subject to some level of uncertainty.
This evaluation consisted of the following tasks:
evaluation of the contribution to various CDF sequences due to the automatic AFW isolation system, assignment of containment failure probabilities, and w aluation of the offsite dose factor, which are presentea below, in this evaluation, the major uncertainties associated with the evaluation of the core damage contribution due to the AFW automatic isolation system are in the assessment of the values for the pertinent individual failures involved in the CDF sequences identified as the most important contributors to risk.
Tl.ese failures are:
1.
Inaovertent actuation of the AFW outomatic isolation systen, 2.
Failure to recover AFW, 3.
Failure to achieve feed and bleeu, and 4.
Fatlure to bypass AFW isolation logic during long-term shutdown.
11
The uncertainties associated with the expected frequencies fcr these events were evaluated by reviewing the uncertainties associated with comparable events involving the AFW and their treatment in References 15, 16, and 17.
It was determined from this review that the error bands for such comparable events were in the range of 5-8.
It was decided that an error factor of 10 would be appropriate for the four events cited above.
One method that could be employed to determine tne uncertainty in the estimated of fsite consequences would be to employ a Monte-Carlo analysis and propagate the distributions through the models.
However, based un statistical cethodology for the log-normal distributions, the combined error factor for the end state of an event tree was approximated to be equal or less than the largest individual error factor of the events used in the estimation of the contribution to CDF (Ref. 18). Hence, an upper bound on the combined error factor is assumed equal to the largest individual error factor.
Uncertainties associated with the probability of containment failure were estimated in a similar manner drawing from the plant-specific PRAs (Ref 4) fiUREG-0933 (Ref. 7), and fiUREG-1150 (Ref. 13),
i The offsite dose release fractions used in the GI 125.!!.7 evaluation were those presented in 14UREG-0933 and WASH-1400 with the exception of Plant C.
The l
fiUREG-0933 f, actions represent the uncertainties of offsite dose calculated for a typical plant. Certain plant-specific characteristics such as assumed source terms and population density surrounding a specific plant introduces scme uncertainty in the calculated offsite consequences.
However, the NUREG-0933 values are considered representative in lieu of a detailed plant-specific 1
evaluation of the offsite consequences. The evaluation of the offsite consequences for Plant C are based on more recent offsite consequence information as part of the NUREG-1150 work.
Based on the previous discussion of the estimated uncertainties, the use i
i nf an error factor of ten was estimated to be representative of the total uncertainties of the factors used to calculate the cost-benefit ratio.
This 1
estimate is reasonable because of the scoping nature of this analysis in order to (1) provide an approximate evaluation of the sensitivity of the recomendation to the uncertainty of tne factors used in the analjsts, and (2) provide an ariditional input to the deterministic considerations for this l
generic issue.
i Table 6 presents the basic information utilized in performing the 4
sensitivity analysis. This table is a compilation of data previcusly j
presented. Table 7 presents the results of the sensitivity analysis.
The 1
sensitivity of the results presented in this table as to the uncertainties in j
the cost-benefit ratios were calculated using an error factor of 10 as describec above.
Cost-benefit ratios were not calculated for those plants j
(Plants 8 and C) for which a net increase in the CDF due to implementing the i
proposed tradifications was estimated. The best estimate cost-benefit ratios for all cases are above the $1,000/perron-rem criterion.
The same is true even I
for the lawer bound estimates with the exception of one for plant A which is
]
$970/ person-rem.
12 4
Taole 6.
Basic Data Used in the Uncertainty Sensitivity Analysis Total Change in C0F Offsite Consequences Cost Cost-Benefit Ratio Plant (perRx-year)
(Total person-rem)
($1000)
(3/ person-rem)
A(CE) 5.5E-07 36 351 9,700 (decrease)
B(BSW) 9.6E-08 4
768 (increase)
BB(B&W) 9.0E.7 44 768 17,200 (decrease)
C (W) 4.0E-07 13 351 (increase)
~
Cost cenefit ratios were not calculated for plants where the implementation of this issue would result in an increase in the estimated risk.
13
A Table 7.
Sensitivity Analysis Results Cost-Benefit Ratio (5/ person-rem)
Based on PRAs (Ref. 4) Con'.ainment Failurc Informattor A (CE)
Upper Buund 97,000 Best Estiniate 9,700 Lower Bound 970 i
BB (B&W)
Upper Bound 172,000 Best Estimate 17,200 Lower Bound 1,700 1
k 6
i i
f i
I 1
I 14 4
~
,w
-.~---,----m,.
,e, n..--,
,.-_,,_,,__.,-..,m.,,.,-
6.
CECISI0ti RAT 10t4 ALE This generic issue was identifica as part of the evaluation of the Davis-Besse event of June 1985 (Ref. 2). Although this issue has a bearing an been under continuing f4RC and industry assessment (Refs.19 and 20) plants ha all PWRs, it was particularly icentified with the B&W plants.
B&W ano one of the focal points of this assessment it directly related to the reliability and operational aspects of the AFW system (known also as Emergency feedwater System (EFS) in B&W terminology).
The recognition of the need to improve the operation of this system has been manifested by the installation of the Emergency Feedwater Initiation and Control (EFIC) System in 3 number of B&W plants, or an acceptable equivalent in other B&W plants.
Based on the results of tiv cost-benefit analyses, and the insights gained during assessment of the pros acd cont rf removing the automatic isolation design feature of the AFW system, we conclude that no backfit requirement to remove this design feature from PWRs is warrantea. This conclusion is based on the following consider 6tions:
1.
The risk reductions realizec in some plants by retaoving the automatic isolation of AFW are very small (on the order of a few person-rem per plant) ano the corresponding cost is in the range of $350,000 to 5800,000 (assuming no replacement power costs and depending on the Hence, both basic backfit criteria (Ref. 21) pump discharge lines).
need to install flow restrictors in the AFW that (1) substantial increase in the overall protection of the public health and safety is achieved, and (2) that the direct and indirect costs of implementation are justified, are not met.
2.
Our assessment of this issue has shown that, for some plants, removal of the automatic isolation of AFW would result in a very small i
increase of the risk to public health and safety.
3.
The installation of the EFIC system in a number of B&W plants has been shown to improve safety and has already been accepted by the staff.
Hence, we recommend that Alternative Resolution tio. 1 "tio Actien" be approved.
7.
IMPLEMEtiTATI0ti I
fio regulatory action is necessary for resolution of this issue. A l
distribution of Reference 4 has been made to include all PWP licensees and venoors. This regulatory endlysis will also de made publicly available as part of its normal distribution, i
i r
w
e 1
REFERENCES 1.
V. S. Nuclear Fegulatory Connission (USilRC), "Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants, LWR Edition,"
NUREG-0800 (formerly issued as i.UREG-75/987), June 1987.
2.
USNRC, "Loss of Hain and Auxiliary Feedwater Event at the Uavis-Besse i
Plant on June 9,1985," NUREG-1154, July 1985.
3.
Memoranduu from H. R. Denton to T. Speis, on Reevaluation Provision to Autocatically Ist, late feedwater from Steam Generator During Line Break, dated September 10, 1986.
4.
Stanley J. Bruske et al., "Evaluation of Generic Issue 125.11.7, Raevaluate Prcvision to Automatically isolate feedwater frou Steam Generator Curing a Line Break," NUREG/CR-5178. EGG-2544, uuly 1988.
5.
USNRC, Office of Inspection and Enforcement, IE Eulletin No. 80-04, t
Analysis of a PWR Main Stream Line Break with Continued Feedwater Addition, February 8,1980.
I 6.
Richard P. Burke et al., "Economic Risks of Nuclear Power Reactor Accidents," NUREG/CR-3673, SAND 84-0178, April 1984 7.
USNRC, Division of Safety Technology, "A Prioritization of Generic Safety Issues;," NUREF 0933, Rev. 2, December 1987.
8.
Hemorandum from Andrew R. Marchese, thru Karl Kniel to Themis Spels, on Specific Person-Rem Calculations With and Without Offsite Protective Actions, dated October 22, 1985.
9.
Memorandum from A. J. Dipalo to G.
D. Hazetis, on Cost Estinates for Generic Issue 125.11.7, dated February 5,1988.
10.
S. W. Heaberlin et al., "A Han<1 book for Value-Impact Assessment,"
huREG/CR-3568, PNL-4646, Decemuer 1983.
11.
J. C. Van Kuiken et al., "Repl6 cement Energy Costs for !!uclear Electricity-Generating Units in the United States:
1987-1991,"
NUREG/CR-4012, Vol. 2. Jenuary 1987, 12.
A. S. Benjamin et al., "Evaluation of Severe Accident Risks and the Potential for Risk Reduction:
Plant C,"
NUREG/CR-4551, Vol. 2 SAfiD86-1309, February 1987.
- 13. USNRC, Office of NVclear Regulatory Research, "Reactor Risk Reference Occument," NUREC-1150, W1.1, Oraf t for Connent, February 1987.
16
o 14.
J. R. Ball, "A liethod for Developing Wick Approximate Estin.ates of Costs for Genaric Actions for liuclear Power Plants," A Handbook for Quick Cost Estimates,fiUREG/CR-4568, AriL/EES-TM-297, April 1986.
- 15. Arthur C. Payne, Jr., Principal Investigator, "Interim Reliability Evaluation Program: Analysis of the Plant A fluclear Power Plant,"
fiUREG/CR-3511/1 of 2, SAf4083-2086/1 of 2, Hay 1984.
l
- 16. Alan D. Sw61n, "Accident Sequence Evaluation Program Human Reliability Analysis Procedure," fiUREG/CR-4/72, SAfiD86-1996, February 1987, 17.
C. D. Gentillion et al., "Generic Data Base and Models" (Chapter on HREP),
EGlG EA-887. June 1982.
18.
S. Bruske et al., "Evaluation of Boiling Water Reactor Water Level Sensing Line Break and Single Failure - Generic Issue 101, Technical Findings and l
Regulatory Analysis (Dratt) " liUREG/CR-5112, June 1988,
- 19. UStiRC, "Safety Evaluation Report Related to Babcock ano Wilcox Owners Group Plant Reassessment Program," fiUREG-1231, flovember 1987.
20.
Ibid, Supplement tio. 1, March 1988.
21.
UStiRC, "Backfit Rule" (10 CFR Part 50.109(a)(3)), Federal Register, Vol.
50, p. 38097, September 20, 1985, e
i 17 1