ML20245B358
| ML20245B358 | |
| Person / Time | |
|---|---|
| Issue date: | 05/31/1989 |
| From: | Basdekas D NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | |
| References | |
| REF-GTECI-115, REF-GTECI-NI, TASK-115, TASK-OR NUREG-1341, NUDOCS 8906230143 | |
| Download: ML20245B358 (26) | |
Text
__
{f NUREG-1341 6
Regulatory Analysis for the Resolution of Generic Issue 115, Enhancement of the Reliability
- f of the Westinghouse Solid State Protection System t
bi
- f U.S. Nuclear Regulatory j
Commission k; '.
Office of Nuclear Regulatory Research D.L. Basdekas pan 9 ou 2gggge p
i
NUREG-1341 Regulatory Analysis for the Resolution of Generic Issue 115, Enhancement of the Reliability of 1:he Westinghouse Solid State Protection System se""";22e"TallM"""" ''
D.L. Basdekas Division of Safety issue Resolution Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555
.s u.
e AVAILABILITY NOTICE Availability of Reference Materials Cited in NRC Publications Most documents cited in NRC publications will be available from one of the following sources:
1.
The NRC Public Document Room 2120 L Street, NW, Lower Level, Washington, DC 20555 2.
The Superintendent of Documents, U.S. Government Printing Office, P.O. Box 37082, Washington, DC 20013-7082 3.
The National Technical Information Sewice, Springfield, VA 22161 Although the listing that follows represents the majority of documents cited in NRC publica-tions, it is not intended to be exhaustive.
Referenced documents available for inspection and copying for a fee from the NRC Public Document Room include NRC correspondence and internal NRC memoranda; NRC Office of Inspection and Enforcement bulletins, circulars, information notices, inspection and investi-gation notices; Licensee Event Reports; vendor reports and correspondence; Commission papers; and applicant and licensee documents and correspondence.
The following documents in the NUREG series are available for purchase from the GPO Sales Program; formal NRC staff and contractor reports, NRC-sponsored conference proceed-ings, and NRC booklets and brochures. Also evallable are Regulatory Guides, NRC regula-tions in the Code of Federal Regulations, and Nuclear Regulatory Commission issuances.
Documents availabio from the National Technical Information Service include NUREG series reports and technical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Nuclear Regulatory Commission.
Documents available from public and special technical libraries include all open literature items, such as books, journal and periodical articles, and transactions. Federal Register notices, federal and state legislation, and congressional reports can usually be obtained from these libraries.
Documents such as theses, dissertations, foreign reports and translations, and non-NRC conference proceedings are available for purchase from the organization sponsoring the publication cited.
Single copies of NRC draft reports are available free, to the extent of supply, upon written request to the Office of information Resources Management, Distribution Section, U.S.
Nuclear Regulatory Commission, Washington, DC 20555.
Copies of industry codes and standards used in a substantive manner in the NRC regulatory process are maintained at the NRC Library,7920 Norfolk Avenue, Bethesda, Maryland, and are available there for reference use by the g ublic. Codes and standards are usually copy-righted and may be purchased from the originating organization or, if they are American National Standards, from the American National Standards Institute,1430 Broadway, New York, NY 10018.
ABSTRACT Generic Issue 115 addresses a concern related to the This regulatory analysis includes a quantitative assess-reliability of the Westinghouse reactor protection system ment of the costs and benefits associated with the various for plants using the Westinghouse Solid State Protection options for enhancing the reliability of the Westinghouse System (SSPS). Several options for improving the SSPS and provides insights for consideration as industry reliability of the Westinghouse reactor trip function for initiatives. No new regulatory requirements are these plants and their effect on core damage frequency proposed.
(CDF) and overall risk were evaluated.
iii NUREG-1341
CONTENTS Abstract....
iii 5.2.3 Utility Engineering and Quality Assurance Costs
.... 7 Preface........
.......... vii g
,,7 Executive Summary......
ix 5.2.5 Training Costs 7
5.2.6 Maintenanc,e Procedures and
- 1. STATEMENT OF PROllLEM.
1 Operating I rocedures Development
- 2. OlUECTIVE..
...1 Costs........
8
- 3. ALTERNATIVE RESOLUTIONS......
1 5.2.7 Recordkeeping and Reporting Costs. 8 5.2.8 Utility Licensing Costs.
.8 3.1 Option 1 - New UV Driver Card......... 2 5.2.9 NRC Review Costs...
.. 8 3.2 Option 2 - Diverse and Redundant New 5.2.10 Averted Onsite Costs........... 8 Relay UV Driver....
......2 5.3 Cost-Benefit Summary..
9 3.3 Option 3 - Overcurrent/ Fusible Link.
...2 5.4 Cost. Benefit Uncertainties.......
.. 9 3.4 Option 4 - Relay logic System.
.2
- 6. DECISION R ATIONALE............
13 3.5 Option 5 - Redundant Shunt Trip Coil
...2 6.1 Option 1 - New UV Driver Card.
. 14 3.6 Option 6 - Contactors....
.2 6.2 Option 2 - Diverse and Redundant New
- 4. TECIINICAL FINDINGS
SUMMARY
.2 Relay UV Driver.....
14 6.3 Option 3 - Overcurrent/ Fusible Link.. 14
- 5. CONSEQUENCES
.3 6.4 Option 4 - Relay Log. System....... 14 ic 5.1 Consequence Analysis
...4 6.5 Option 5 - Redundant Shunt Trip Coil.
15 5.2 Cost Analysis...
,. 5 6.6 Option 6 - Contactors..
......... 15 5.2.1 Replacement Power Cost 7
- 7. IMPLEMENTATION................. 15 5.2.2 Hardware Corts.........
..7 REFERENCES
......... 16 TABLES 1 Core damage frequency results 3
7 Cost-benefit including averted onsite costs
($/ person-rem reduction)......
. 10 2 Consequence analysis data 4
8 Base case risk uncertainty results 11 3 Consequence analysis results...
5 (person. rem)
.~.......... -.....
9 Cost-benefit uncertainty results.....,.... 11 4 Options cost analysis data (cost $K)...
6 10 Point estimate / uncertainty results 5 Averted onsite costs ($)..
.8 comparison..
...... 13 6 Cost-benefit summary without averted onsite 11 Option 6 versus Option 3 cost. benefit costs ($/ person-rem reduction)....
9 uncertainty comparison.
........ 13 FIGURES 1
Cost-benefit uncertainty distributions.
12 v
PREFACE This report presents the regulatorv analysis, including the and other insights gained during this effort have shown decision rationale, for the resolution of Generic Issue that no new regulatory requirements are warranted in ac-115. The objective of this regulatory analysis is to deter-cordance with the backfit rule,10 CFR Part 50.109(a)(3).
mine whether the options evaluated as part of this effort Certain insights may be considered by the NRC and licen-warrant implementation under the backfit and A'IWS secs / applicants of the affected plants for possible industry rules. The risk change estimates, cost-benefit analyses, initiatives.
Demetrios L.13asdekas vii NURIiG-i341
EXECUTIVE
SUMMARY
'Ihis report provides supporting information, including a This regulatory analysis is partially based on a reliability value-impact analysis, for the Nuclear Regulatory Com-analysis performed at the Idaho National Engineering mission's (NRC's) resolution of Generic issue 115, En-12boratory(INEL)of the reactor tripactuation,whichin-hancement of the Reliability of the Westinghouse Solid cludes all the system and plant modifications required in State Protection System. Generic Issue ll5 addresses the accordance with existing regula'. ions and the KnVS concern that certain parts of the reactor trip system,in-rulemaking requirements. These modifications consist of cluding the reactor trip breakers (RTils), may be contrib-the automatic actuation of the shunt trip device on the uting to an unreliability level of the reactor trip function RTils and the AMS AC modifications. This configuration resulting in significant levels of risk. It was proposed that was taken to be the base case configuration for this analy-this risk might be mitigated by one or more design changes sis.
identified by the NRC staff as potential options in en-hancing the reactor trip function reliability in Westing.
Ilecause the reliability of the RTIls is an important cle-house (W) plants using the Solid State Protection System ment of the overall resetor trip reliability, it was also as-(SSPS).
sessed as part c.f this effort. Relevant data from opera-tior.a! experience, including the Nuclear Plant Reliability Generic Issue 115 was identified during the anticipated Data System (NPRDS) data, were evaluated and incorpo-transients without scram (ATWS) rulemaking decision rated as appropriate in the calculation of the failure rates process.The final ATWS rule (10 CFR 50.62) required a f such major system components as the R'llis.
diverse scram system for Combustion Engineering (CE) and llabcock & Wilcox (ll&W) plants, but not for the W
,N anah m. resuhs show a current level of reliability plants. 'Ihe ATWS r ulemaking analysis did not show a Ig comparable with previous regulatory analyses such as vorable cost-benefit ratio for inclusion of both a diverse those m the Amendments to 10 CFR Part 50 Related to scram system and diverse auxiliary feedwater initiation KnVS Events (SEL,Y-83-293) and thc 6cnenc implica-tions of Kn; S Events at the Salem Nuclear Power Plant and turbine trip actuation for E plants. Therefore, the W plants were required only to have the diverse auxillag (Generic Letter 83-28).The rcsults also take mto account feedwater initiation and turbine trip actuation, known as the currently required regulatory modifications. Since the KnVS Mitigating Systems Actuation Circuitry (AM-there is little expen,ence to date of operation of the reac-SAC) tor trip system with the automatic shunt trip, an estimate of the improved reliability of the system was performed by modeling the system and performing reliability calcula-As a consequence of the Salem KnVS cvents in 1983, tions using the Integrated Reliability and Risk Analysis Genenc Letter 83-28 established the requirement for the System (IRRAS) fault tree computer program. The re-automatic actuation of the shunt trip attachment of the sults of this modeling show an improved reliability in line R l'Ils forEand il&W plants (this feature was meluded m with that estimated by the KnVS Task Force, which rec-the original design for CE plants). Although this feature ommended the modification.
provides additional diversity and increased reliability of Ihe reactor trip system, the reliability of the automatic ac-Results of the reactor trip reliability analysis were applied tuation signal to the RTIls and the RTHs themselves still to a generic W ATWS event tree model. The base case had to be considered.
~
core damage frcquency due to an KnVS cvent sequence was calculated to be 5.0E-06/Rx-year. Six options for im-l The E SSPS design uses a solid state board assembly, proving the reliability of the reactor trip system were called the undervoltage (UV) driver card, to actuate the identified. The changes in core damage frequency (CDF)
UV trip and the shunt trip. The potential common cause between each of these options and the base case system l
failure (CCF) of the UV driver card was evaluated in the were calculated. The reductions range from 1.0E-07 to Generic issue 115 prioritization analysis, and it was deter-2.lE-06 events /Rx-year. One option resulted in an in-mined to be the dominant contributor to failure of the re-crecsc in risk.
actor trip signal to the RTils. More recent individual fail-urcs of thc E UV driver card raised the question whether The changes in CDF between the base case and the op-there might be a higher probability of SSPS failure than tions were applied uniformly to a consequence analysis that calculated during the KnVS rulemaking proceeding.
and cost-benefit analysis using both point estimates and A higher failure rate of the UV driver card would lead to a probability distribution mean estimates of risk and costs.
higher probability of CCF of both reactor trip signal The results of the cost-benefit analysis for the options trains. A higher CCF contribution, in turn, would in-were calculated with and without the potential for manual crease the risk due to offsite consequences of an KnVS operator scram, which must take place within one minute event.
from the time of receipt of a reactor trip signal to avoid ix NUREG-1341
core damage in worst case scenarios.The results, without taking into accouit averted onsite costs, are shown below:
Cost. Benefit
($/ Person. Rem Reduction)
Cost $K Without Operator With Operator Option CDF per Plant Action Action 1
1.0E-7 50 470 7.750 2
2.0E-7 81 616 10,048 3
2.1 E-6 132 843 981 4
4.0E-7 1,084 3,250 48,299 5
-9.0E-7 201 No benefit No benefit 6
2.1E-6 243 1,120 1,772 When averted onsite costs are taken into consideration, the following results are obtained:
Cost.llenefit
($/ Person. Rem Reduction)
Net Cost $K Without Operator With Operator Option per Plant Action Action 1
49 320 7,620 2
80 464 9,890 3
112 689 833 4
1,080 3.100 48,100 5
190 No benefit No benefit 6
223 1,063 1,626 l
An uncertaimy analysis was performed using probability are warranted in accordance with the backfit rule,10 CFR distributions for risk and costs.The results for the mean Part 50.109(a)(3).
cost-benefit estimates based on these distributions are all higher when compared with the point estimates (i.e.,less Based on insights gained during the technical evaluation cost-effective) and substantially higher than the nominal of this issue, we believe that sufficient incentives exist for screening value of $1,000/ person-rem.
subsequent initiatives by the affected licensees and appli-cants. Such incentives, although not regulatory require-Based on the results of the cost-benefit analyses and the ments, would contribute toward reducing the regulatory insights gained during the evaluation of the six options to burden on licensees, while encouraging enhanced reli-enhance the reliability of the Westinghouse SSPS, as well ability of the reactor trip function.The details of these in-as the relatively small risk reduction provided by the ana-sights are further developed in the regulatory analysis lyzed options, we conclude that io backfit requirements (Section 6).
REGULATORY ANALYSIS FOR TIIE RESOLUTION OF GENERIC ISSUE 115, ENIIANCEN1ENT OF TIIE RELIAlllLITY OF TIIE WESTINGHOUSE SOLID STATE PROTECTION SYSTEN1 L STATEN1ENT OF PROllLEN1 SSPS trip function were considered as part of this work as discussed in Section 3 and in N URiiG/CR-5197 (Ref. 4).
The KIWS rule (Ref.1)for Westinghouse (E) plants re.
In an analysis performed by PNL (Ref. 5), it is assumed quired the implementation of a diverse ATWS Mitigating that the AMSAC required by the KlWS rule for E plants Systems Actuation Circuitry (AMSAC). The functions is in place and operational. llased on this analysis, a prescribed for AMS AC are automatic turbine trip and the prioritization was performed (Ref. 6) resulting in the as-initiation of auxiliary feedwater, independent of the reac-signment of high priority to this issue. It was designated Generic issue 115, "linhancement of the Reliability of tor trip system.
Westinghouse Solid State Protection System."
As a consequence of the Salem KlWS events of February Ilaving investigated five UV driver card failures, ~W deter-1983 Generic Letter 83-28 (Ref. 2) established the re-mined that these failures were caused by poor nainte-quirement that the shunt trip attachment of the reactor nance and test-related practices.These practices involved trip breakers (RTils) for E and liabcock and Wilcox the inadvertent shorting of the scram breaker's UV trip (ll&W) plants should be provided with automatic actua-co I, causing a shorted failure of the output transistor in tion in addition to the existing manual actuation (this fea-the UV driver card.To climinate this safety problem,W ture was iacluded in the original design for Combustion modified the design of the UV card to provide a fusitic IIngineering (Cli) plants). Although this modification link in the output circuit that will open the circuit when provides a significant increase in the reliability of the the UV coil is shorted.The modification will produce a RTils and, hence, the reactor trip system, it had not been UV trip signal to the respective RTil that will persist until previously pursued as an action to reduce the potential of the card is removed, repaired (by E), and replaced. Wes-an NIWS event during the extensive dialogue and study tinghouse Technical Ilulletin NSID-T8-85-16 (Ref. 7),
of the NIWS issue. Indeed, certain other options for in-dated July 31,1985, was issued to the W utilities, as re-creasing the reliability of the reactor trip system for E quired by the Salem KlWS Generic Letter 83-28 plants had also not received detailed consideration.
(Ref. 2), recommending installation of modificd UV driver cards. The Westinghouse Technical llulletin also Westinghouse plants with the Solid State Protection Sys-recommends specific maint enance and test procedures to tem (SSPS) design have had recent failures of the under-be followed to prevent f adures of this type pending instal-voltage (UV) driver card that have raised concerns about lation of the modified UV driver card. An informal survey the susceptibility of the design to common mode and ran-of the affected plants by the NRC staff indicates that at dom failures of redundant components. I!nhancement of least seven of the 37 plants have installed the new UV the reliability of the E SSPS was discussed by NRR in a driver cards and modified their test procedures, or they 1985 request for consideration (Ref. 3).
are in the process of doing so. Most of the remaining 30 plants have adop;cd only the new test procedures.
The observed failures of the UV driver card suggest a higher probability of SSPS failure than that calculated
- 2. OIUECTIVE during the NIWS rulemaking proceeding. A higher prob-ability of SSPS failure in turn would lead to a higher prob.
The objective of this regulatory analysis is to determine ability of NI'WS and, consequcally, to a higher risk to the whether any of the six options intended to enhance the offsite population surrounding E plants equipped with reliability of the E SSPS are warranted under existing re-the SSPS. These systems are used in 30 operating g quirements, including the backfit rule.
plants and seven under construction.
- 3. ALTERNATIVE RESOLUTIONS Incorporation of additional diversity for the UV diiver
.!here were two basic alternatives (Refs. 6 and 8) consid-card function would reduce the probabili:y of, an Al% S cred as a basis for resolution of Generic Iscue 115.
event. I or example, on e way that the UV driver reliability could be improved is by installing a relay driver and associ-Alternative 1 ated relays to duplicMe the function of the UV driver, thereby providing diversity for this function. A number of Take no action. Under this alternative there will be no additional options to improve the reliability of the E new regulatory requirements. Consistent with existing i
NURiiG-1341
regulations, this alternative does not preclude a licensec, would receive the same input from the analog instrumen-or an applicant for an operating license, from proposing tation input relays as the SSPS logic and provide a redur.-
to the NRC staff design changes intended to enhance the dant UV trip output to the RTils.
reliability / operability of the reactor trip system and its components on a plant-specific basis, nor does it preclude 3.5 O tion 5 - Redundant Shunt TriE E
E, w hich is the NSSS and R FIl vendor, flon; doing so on a generic basis.
Coil
'lhis option, in replacing the UV trip mechanisms on each Alternative 2 RTil with a shunt trip mechanism, provides a redundant shunt trip configuration for each RTil (Ref. 9)/lhe shunt Require design changes corresponding to one or more of trip de power supply would be supplemented with a fail-the six options to enhance the reliability of the E SSPS safe capacitor circuit that would supply a de charge to op-that were evaluated under this effort (Ref. 4). The six op-crate the shunt trip on loss of the de power supply.
tions, proposed by E or the NRC staff, are described briefly below.
3.6 Option 6 - Contactors 3.1 Option 1 - New UV Driver Card Under this option, one RTil and its corresponding bypass breaker would be replaced each by a contactor (Ref.10).
'Ihis option entails the replacement of the UV drivercard with one that was modified and recommended by ~W (Ref. 7). E modified the design of t he UV card to provi le
- 4. TECHNICAL FINDINGS a fusible link ir the outpet circuit that will open the circuit
SUMMARY
when the UV coil is shorted. The modification will pro-duce a UV trip signal to the respective RTil that will per-sist until the card is removed, repaired (by E), and re-Probabilistic methods were used for a generic assessment placed. The new UV driver card with a fuse on the 48-volt of the changes in risk that would occur for each of the six output would have prevented four of the five short-circuit options to augment the reliability of the basic design of failures expenenced by the onginal UV driver card de-the reactor trip system in W plants using the SSPS. The
~
SI "'
method and results of this assessment by the Idaho Na-E tional lingineering l2boratory (INEL) are reported in NURI!G/CR-5197 (Ref. 4). Ilasically, the assessment 3.2 Option 2 - Diverse and Redundant was performed by constructing a reliability model of the New Rela}* UV Driver ator nie syuem and qu nMyin8 the auendant fauh trees for the reactor tnp base case and its modified design This option provides a diverse and redundant UV tnp reflecting each of the six options described in Section 3, path consisting of relays arranged in parallel with the UV using the IRRAS (Ref. I1) computer program.
driver card. These relays would receive the same reactor trip inputs from the SSPS logic cards and provide driver The core damage frequency (CDF) calculations were per-relays for the RTil's UV trip and shunt trip devices (Refs.
formed using the generic ATWS cvent trees developed 6 and 8), basically duplicating the function of the UV for E plants as part of the ATWS rulemaking regulatory driver card by diverse means.
analysis (Ref.12). The event trees for E plants were de-veloped for turbine trip initiating events and nonturbine trip initiating events. These event trees were quantified l
3.3 Option 3 - Overcurrent/ Fusible with the AMSAC modifications included (Refs.1 and 12).
Link l
The early ATWS sequence is a very short-term sequence This option entails the incorporation of design changes /
with potentially high consequences. The sequence starts additions to provide a diverse tripping mechanism redun-with a transient-initiating event on the secondary side of a Emt and diverse to the RTHs that would create an over-PWR plant. ATWS events taking place at low power lev-current condition on receipt of a reactor trip signal, els (< 25%) have been judged to be inconsequential to the causing the opening of a fusible link placed in series with pla*. Failures of additional safety systems would be re-the RTils.
,,,i.ed to cause core damage. The reactor system design parameters that most af feet the NIWS sequence are the 3.4 Option 4 - Relay Logic System pnsurim to n actor coolant volony ratio and prunary system pressure relief capacity. In A plants these two pa-This option provides a diverse trip logic composed of a re-rameters are more favorable toward KlWS mitigation lay logic system in parallel with the SSPS reactor t rip logic than in plants supplied by the other PWR vendors. Also, and respective UV driver cards. This diverse logic system these two parameters are nearly the same in three-and NUREG-1341 2
four-loop E plant designs, which makes all E SSPS slower as the temperature riscs.The slower the reaction plants similar for A'IWS analysis purposes.
shutdown, the higher the pressure transient in the reactor coolant system, which gives a higher potential for core Another variable affecting the ATWS pressure transient damage.
is the moderator temperature coefficient (MTC). The MTC is the measure of how reactor power (reactivity) var-The mean ATWS core damage frequencies (CDFs) are ics with moderator (rcactor coolant) temperature. If the shown in Table 1 for two cases, with and without operator MTC is more negative, then the fission reaction is shut action to manually scram the reactor. An uncertainty down faster as the moderator temperature rises. If the analysis using Monte Carlo simulation was performed by MTC is less negative, then the reaction shutdown is INI!L (Ref. 4) and is discussed in Section 5.4.
Table 1. Core damage frequency results Core Damage Frequencies" (Events / Reactor Year) llase Case Option 1 Option 2 Option 3 Option 4 Option 5 Option 6 Without operator action 9.9E-6 8.3E-6 7.911-6 7.5E-6 4.8E-6 1.11!-5
- 8. l E-6 to manually scram Change in core damage 1.6E-6 2.0E-6 2.4 11-6 5.11!- 6
-1.1 E-6 1.8E-6 frequency With operator action 5.011-6 4.9E-6 4.81i-6 2.9 11-6 4.6E-6 5.9 E-6 2.91!-6 to manually scram Change in core damage 1.0E-7 2.0E-7 2.111-6 4.011-7
-9.0 E-7 2.111-6 frequency For comparison, the NDVS goal ntabhshed by the NRC is 1.01L5 cvents/ reactor year (Ref.12).
Without the operator action, the results show the contri-
- 5. CONSEQUENCES bution to a change in CDF due only to the automatic reac-tor trip function. With the operator action, the results This section assesses the cost-benefit aspects of the alter-show the contribution to change m LDF varies with both native resolutions of this issue.
the manual operator reactor trip and automatic reactor trip functions.
In such an assessment," costs" provide a measure of pri-marily economic consequences resulting from the imple-Without operator action taken into account, the CDF re-mentation of alternative resolutions. Based on their defi-sults in Table I show only slight improvements. For Op-nition, these costs may be considered positive (e.g., the tions 1, 2, 3, and 6, the decrease in CDF is less than incurred costs in 1988 U.S. dollars for installing, operat-3.0E-06 events /Rx-year. For Option 4, the CDF de-ing, and maintaining the plant modifications needed to crease is 5.11!-06 events /Rx-year, and for Option 5, the implement a resolution, including the cost of any replace-I CDF increases.
ment power during a necessary downtime for the plant),
l or they may be considered negative (e.g., savings to the op ng y n ums o a a
ent mus ass With the operator action taken into account, the improve-ated wjth plant apaks, deamip, power mplamment, ment options are even less effective because the operator etc.). 'l hus, the net cost represents the positive costs mi-can mitigate many automatic reactor trip failures with a pawnt wod of the nega% costs (amted on-nus manual trip action. For Options 1,2, and 4, the decrease site costs) over the remaining lifetime of the plant (s).
in CDF is less than 4.0E-07. For Options 3 and 6, the CDF decrease is 2.lE -06, and for Option 5, the CDF in-Conversely, the term " benefits" denotes the improve-
- "
- S' ments made to public health and safety as measured in the reduction of person-rems of population exposure as As expected, without the operator action included the well as in the reduction of offsite property damage costs options show a greater improvement in CDF and greater associated with land interdiction and decontamination improvement in automatic reactor trip reliability, that may be necessary. In the analysis used for this study, 3
NURIiG-1341
1 l
1 offsite property damage was not estimated separately. In-Methodology Application Program) study to derive a stead, a reasonably conservative allowance in the public person-rem dose with a generic plant site calculation.The dose estimate was used as an adequate surrogate in this only KIWS dominant risk sequence (fpKMU) in this case.
study is assumed to result in a Category 3 release with a conditional probability of 0.5, a Category 5 release with a The number and irnportance of parameters contributing conditional probability of 0.007, and a Category 7 release to the costs and benefits vary with the type of accident and with a conditional probability of 0.5.Thus, based upon the the plant location (Ref.13). In the analyses performed in generic site and the above release categories, which are this study, the site chameteristics for a " typical" midwest.
defined in the PRA Procedures Guide (Ref.15). a ern plant and site (Refs. 4 and M) are used as representa.
weighted average of 2.711 + 06 person-rem /cVent was tive of the population of U.S. PWRs affected by thisissue.
used for the consequences of KlWS events using CRAC code results of the NUREG-0933 analysis.
5.1 Consequence Analysis Two other sources of consequence data were used in the The consequence analysis obtained the person-rem dose evaluation of uncertainties. The first source is from three sources of previous analyses shown in Table 2.
WASH-1400 (Ref.14), from which a consequence of This person-rem dose is based upon an average derived LOE + 07 permn-rem per event yvas used for all core dam-from the consequence analyses of the source studies.
age events.,Ihe second source is Ref erence 16. Two W These source studies derived the person-rem doses per plants Sequoyah and Surry, were evaluated for conse-event based upon various consequence methodologies quence results. The results are shown m Table 2.
that include an evaluation of core damage, source term, and potential release to the environment. A mean (point The 30 operating W plants using the SSPS (see Table 1) estimate) value of 2.711 + 06 person rem per event was have a total of 737 reactor years of operation remaining.
obtained from NUREG-0933 (Ref. 8). NUREG-0933 The calculated reduction of core damage frequency is ap-used the Oconce 3 RSSMAP (Reactor Safety Study plied in the following manner:
CDF Offsite Radiation Reactor Years Total Risk Reduction x
Dose x
Remaining Reduction
=
(livents/ Reactor Year)
(Person-Rem /livent)
(30 Plants)
(Person-Rem)
Table 2. Consequence analysis data Person Rem Source Sequence (s) per Event Reference 14 (WASil-1400) 1.011 + 7 Reference 8 (NUREG-0933) 2.7E + 6 Reference 16 Sequoyah Transient, fail to SCR AM, IIPI fails, or MTC unfavorable 8.7 E + 4 Surry Transient, fail to SCRAM, IIPI fails, or MTC unfavorable 8.0E + 4 Mean 5th 95th Consequence Distribution (1 og-normal) 2.7Fi + 6 1.0E + 6 1.0l! + 7 The CDi' value is in terms of events per reactor year, in terms of events per reactoryear of operation, which re-which reflects the units in the results of the event tree flects the utili/ation factor of the entire population of analysis.The event tree initiating-event data are collected PWRs in the initia:ing-event data base.
The mean (point estimate) value results used in the con-and justification for Table 4 are explained in this subsec-sequence analysis are shown in Table 3. The uncertainty tion.
analysis results are discussed in Section 5.4. The person-rem reductions are the difference between the base case in general, cost estimates are based first on figures de consequences and the options consequences.
rived from referenced sources and from cost estimating methodologies in Generic Cost Estimates (Ref.18). The cost estimates are based on low, best estimate, and high 5.2 Cost Analysis values to represent the uncertainty range of each cost item. The best estimate is used as the mean (point esti-The guidance given in iteference 17 was used to estimate mate) value. In selected cases where costing information and analyze the costs of implementing the six options.
was available on specific components and attendant engi-The results of the cost analysis are presented in Tab!c 4.
neering and regulatory effort, that information was used The definition of cost categories, cost estimate rationale, (Itef.19).
Table 3. Consequence analysis results Core Damage Iteactor Years Total Itisk Frequency Olisite Dose Remaining Total Itisk Iteduction SSPS (Esent/Itcactor Year)
(Person. Item /Esent)
(for 30 Plants)" (Person Item)
(Person Item)
Without Operator Action flase Case 9.91!-06 2.7E + 06 737 19,614 Option 1 8.3E-06 2.7E + 06 737 16,425
-3,189 Option 2 7.9E-06 2.7E + 06 737 15.670
-3,944 Option 3 7.51!-06 2.711 + 06 737 14,914
-4,700 Option 4 4.811-06 2.71! + 06 737 9,614
-10,000 Option 5 1.111-0 5 2.711 + 06 737 21,473 1,869 Option 6 8.1I!-06 2.711+ 06 737 16,118
-3.496 With Operator Action llase Case 5.Oli-06 2.711 + 06 737 9,882 Option 1 4.911-06 2.7I1 + 06 737 9,689
-193 Option 2 4.811-06 2.711 + 06 737 9.640
-242 Option 3 2.911-06 2.71? + 06 737 5,846
-4,036 Option 4 4.611-06 2.711 + 06 737 9,209
-673 Option 5 5.911-06 2.7E + 06 737 11,685 1,803 Option 6 2.91i-06 2.711 + 06 737 5,770
-4.112
~
llased on 3n-year plant Ide multiplied by 30 plants. equah 90n total reactor years minus 163 reactor years of cumulative operating experience of a
the 30 plants.
5 NURIIG-1341
h 0
3 ig 7
6 1
9, 8
7 5
1 5
2 5
3 5
l f
1 2
0 1
0 4
4 2
4 3
5 1
1 9
6 n
tea o
tem 0
s i
tp Bi 9
3 t
4 0
5 6
6 0
2 5
4 2,
O sE 0
8 6
1 0
3 3
2 3
2 2
7 w
00 o
0 2
0 3
7 1
0 6
3 3
9, l
0 4
3 1
0 2
3 1
1 1
1 3
0 h
5 g
2, 7
5 5
ii l
18 4
0 5
1 2
5 7
14 1
4 7
0 4
4 9
4 3
4 5
n tea o
tem 0
s i
tp Bit 5
0 6
6 2
2 6
3 s
1 0,
O 2
5 0
E 0
9 3
5 0
3 3
7 w
00 o
0 7
0 7
1 6
3 1
0 4
2 3
0 2
3 5
1 1
1 3,
l 3
0 h
0 6
8 ig 5
9 8,
4 2,
5 1
5 2
5 2
i 3
9 1
8 l
0 4
4 2
4 3
2 6
)
4 te K
n a
S t
io tem 0
4 20 s
Bit 0
8 5,
s tp c
O E
0 10, 6
6 0
2 5
0 2
s o
0 3
3 2
3 2
1 3
(
a 0
t w
5 a
o 0
5 8,
d 7
1 0
6 3
9 3
s 0
t 5
7 0
2 3
1 1
1 7
2 is Q
0 ly 7
a 7
9 3,
8 2
0 2
1 1
2 5
7 I
n I
I a
1 3
6 1
0 1
1 9
4 3
3 1
t 3
o n
te s
c o
t a
m 0
s s
itp Bit 6
2 5
3 e
6 n
2 i
O E
0 1
0 o
9 9
9, s
o 4
t 0
0 0
7 3
2 1
3 tp w
0 O
6 o
0 8
8 6
3 4
0 8
2 8
0 0
0 5
1 1
2 1,
l 7
2 le h
0 b
ig 7
8 4
R 3
5 1
1 2
5 9
9 a
i l
1 6
3 0
1 1
9 4
3 2
8 T
l 2
te n
a t
io s
m 0
e Bit 5
5 9
9 3
tp 2
5 1
4, s
O E
0 8
4 2
0 0
0 7
3 2
8 2
w 09 e
5 8
8 2,
6 3
3 l
0 3
3 1
0 0
0 5
1 1
4 1
h_
0 ig 8
6 i
4 5
5 1
1 0
o 6
9, l
9 3
7 0
0 1
1 7
1 t
1 4
1 te n
a o
t m
0 se it Bi 0
p t
8 2
9 9
O E
0 2
5 0
0 0
0 5
5 5
5 15, s
0 w
0 o
8 8
0, l
1 0
0 0
0 3
3 3
3 1
5 0
2 3
g g
s n n ie tnt g
e ipi c s s
t w
n n
r es n e e
er
)
o e
i o
r g r o
r r
e o g
ie t
)
mc a
eA i
a u n u s
g n
v n
k p is a
yw e
t g
nd id y
t e
e e
e n
ta a
tynQ la r
o e o od ity r
la tr lal e
r i
te n
e t
e d
c r
n n c a
C lpo tod i igd la ia i r pr cn li ic R
to nT 3 c
r ne C
ls p s
w r
l ts a
r ut p a r c a inn p
e a t
e d o0 e p ah t
ea n
r R
F U
I T Mp O R
Ul N T (p I
i 2EC i
5.2.1 Replaecment Power Cost A cost estimate for Option 3 was calculated and provided in Reference 19. The best estimate for factory hardware The replacement power cost category includes any cost for Option 3 was $16,000. The low estimate was based on related to plant downtime owing to the required plant half (58,000), and the high estimate was based on double modifications. This downtime is a result of installing the
($32,000).
modificatitm or is a result of inadvertent plant shutdowns caused by the modification after installation.Thisanalysis A cost estimate for Option 4 was taken from the KlWS assumed all options could be installed during plant rulemaking analysis (iter.12).The cost of a diverse reac-refueling shutdowns. The only contributor to replace-tor trip system for W plants was estimated at $1 million.
ment power costs is inadvertent shutdowns due to the This includes hardware, utility engineering, and installa-modification installed. Options 1,2,3,5, and 6 require in-tion. Therefore, the $1 million cost total is applied to cost significant equipment additions. Option 4 requires a rela-Categories 2,3, and 4 for Option 4.
tively large amount of key equipment for a diverse logic and actuation system and has a greater potential for inad-In assessing the cost of Option 5 the required modifica-vertent trips. Ilowever, Option 4 can be properly engi-tions of four itTils and two spares (six total) were neered, installed, and maintained, such that it should not estimated at $ 10,000 each. In addition to the itTil modifi-cause an appreciable merease in inadvertent trips.There-cations, some relays, two capacitor circuits, and miscella-fore, the best estimate increase m. advertent shutdowns is neous equipment would be required at an estimated cost negligible, if any.
of $35,000. The total for this option's hardware is $95,000 as the best estimate, with half ($47,000) as the low esti-mate, and one and a half ($ 145,000) as the high estimate.
liven though the best estimate increase in inadvertent trips is basically zero for all options, there is some change in the probability of inadvertent trips, and it is accounted 5.2.3 Utility Engineering and Quality for in the high-estimate numbers of the distnbution used Assurance Costs in the unecrtainty analysis presented in Section 5.4. liach reactor trip is assumed lo cause a half day outage or to cost The cost of utility engineering and quality assurance is
$250,000 per trip. The statistical data for the number of based on an engineering hourly cost of $100/ hour (llef.
trips per reactor year of operation were obtained from 18). The hours for each option are 50,45,400,300, and iteference 20. The high-estimate incremental increases 600 for Options 1,2,3,5, and 6, respectively (these costs in the number of inadvertent trips over the remainirg hfc=
for Option 4 are included in Section 5.2.2). The estimate time of the population of plants are 0.37,0,75,0.75,3.73, for Option 3 was provided in iteference 19, and the other 0.75, and 0.75 events for Options 1 through 6, respec-estimates were made taking into account the complexity tively.
of Option 3.The low and high estimates were based on a difference from the best estimate of $2,000, $1,500,
$20,000, $ 10.000, and 30.000, respectivciv, for each of the
.lhe number of events for each option multiplied by other options.
~
$250,000 per event is equal to the cost attributed to these events and is used as the high-cost estimate for inadver-tent trip downtir.ie.
5.2.4 Installation Costs Installation costs are assessed at a labor rate of $25/ hour (itef.18). Ilours estimated for installation oi cach of the 5.2.2 Ilardware Costs options are 8,100,400. and 200, respectively. for options 1,2,3, and 5 (these costs for Option 4 are included in Sec-This category includes the cost of all required hardware tion 5.2.2).,ihe best estimate costs are $200, $2,500, for the modifications for each option.The cost of two re-
$ 10,000. and $5,000.1he low cost for Option 1 is zero with placement and two spare (four total) UV dnver cards and a @ et of $500.1,or Option 2, the low and high esti-technical support was determined to be $28.000. This fig-man wm a mna kom the best estimate.
l ure was used as the best estimate, with the cost of three r Options 3 and 5, the low and high estimates were a cards ($21,000) for a low estimate, and a 20 percent in-dMmna h, om th M Mmat crease in cost as the high estimate.
aining Gsts l'or Option 2, the factory hardware for relays and other hardware is estimated to be $3.000 (iter. 8). An estimate After careful consideration of the nature of modifications for a qualified enclosure in addition to the relays and involved in all six options, it was concluded that, because other hardware adds another $5,000. Therefore,5S,000 is of thesimilarityand/orsimplicityof the equipment, main-used as the best estimate, $3,000 as a low estimate and an tenance. and testing practices, no training would be re-additional $5.000 for the high estimate of $13,000.
quired.
7 NUltliG-1341
5.2.6 Maintenance Procedures and 5.2.9 NRC Review Costs Operating Procedures Development The NRC review cost estimate is taken from Reference 8.
Costs The best estimate given in this analysis is approximately
$25,000.The low was estimated at $ 13,000 and the high at Procedures will have to be written for all option modifica-
$35,000. Option 1 used a lower estimate, based on lesser tions. Estimates for procedure changes for two catego-complexity, w hich is the sr.mc as the utility licensing costs ries, routine and complex, are provided in Reference 18.
for Option 1.
The estimates are given as best, low, and high estimater for each of the categories. Options 1,2, and 3 would re-5.2.10 Averted Onsite Costs quire routine procedure changes, while Options 4,5, and In addition to the cost associated with the modifications, 6 would require complex procedure changes.
averted onsite costs can have a significant effect on the overall cost-benefit ratio depending on the expected re-5.2.7 Recordkeeping and Reporting Costs duction in CDF. They include cleanup, repair, and re-placement power costs. These costs may be estimated by The initial cost for evaluating changes in recordkeeping multiplying the change in accident frequency by the dis-and reporting requirements is approximately $40,000 for counted onsite damage costs. Ilased on the guidance of a complex change and ranges downward from $40.000 for Reference 22, an estimate of these costs on a per plant changes of less complexity (Ref.18). After the initial cost, basis was made for each option using the following equa-it is assumed that the continuing cost is negligible, and the tion:
additional reporting and recordkeepireg can be accom-where Vop= N AFU plished by the current staffing. The best estimates for these costs are based on the complexity of the option. Op-V value of onsite averted costs
=
tion 1 is considered to be the least complex, with Options oI3 number of affected facilities 2,3, and 5 more complex, and Options 4 and 6 the most N
=
complex. The low and high estimates viere taken as a AF reduction in accident frequency
=
$2,000 difference from the best estimates for Options 1, present value of onsite property damage U
=
2,3, and 5, and a $ 10,000 and $5,000 difference for t he low and high of Options 4 and 6.
E..sumated values for U, present value of onsite property
"*"E"'""""
5.2.8 Utility Licensing Costs Ilest estimate
$1.011+ 10/ severe accident event The utility licensing costs are based on Reference 21, a cost analysis for Generie issue 125.11.7, which showed a 1ligh estimate
$3.0E + 10/ severe accident event utility bcensing cost of $32,000 per plant for a complex low estimate
$7.Oli+ 9/ severe accident event modification. Options 2,3,4,5, and 6 are assumed to be complex changes with $32.000 as the best estimate. The These values are multiplied with the expected change in low and high estimates are based on a differcnce of CDF for each option shown in Table 1.The results of the
$ 16,000 and $ 10,000. Option 1 is of lesser complexity, so a analysis for a 10 percent discount rate are shown in Table lower estimate ($5,000) and high and low differentials are
- 5. If a 5 percent discount rate is assumed, the averted on-applied.
site costs would be twice as high.
Table 5. Averted opsite costs ($)
Without Operator Action With Operator Action low Ilest fligh low Hest liigh Option 1 11.200 16,000 48.000 679 970 2,400 Option 2 14.000 20,000 60,000 840 1,200 3,600 Option 3 16.800 24,000 72.000 14,000 20,000 60,000 Option 4 35,000 50,000 150,000 2,380 3.400 10,200 Option 5 6,510 4,300 27,900 6,370 9.100 27,300 Option 6 12,600 18.000 54,000 14,000 29,000 60,000 NURiiG-1341 S
5.3 Cost-Ilenefit Summary quately represent the E reactor trip system and generic W ATWS sequence for the quantifications and compari-
~$ns used in this analysis. This uncertainty evaluation ad-The cost-benefit analysis was performed in accordance s
with the following equation:
dresses the uncertainty associated with the data used in Person-Rem Cost /Henefit Option Cost + Reduction 7he basic approach to the evaluation of the cost-benefit
=
(30 Plants)
(30 Plants)
($/ Person-Rem)
- t. uncertainties uses a Monte Carlo sampling program called @ RISK (Ref. 23). The simulation program was
,lhe mean (point estimate)value results of the cost-bene-used to evaluate the uncertainty of the cost-benefit, rep-fit analysis are shown m fable 6. The results m rable 6 resented by cost distributions and risk reduction distribu-summarize only the costs listed in Table 4.The costs asso-tions. The cost distributions are represented by a triangu-ciated with averted onsite costs (Subsection 5.2.10) are lar distribution with the low and high cost estimates not included in the cost-benefit results ofTable 6. Ihc im-forming the base of the triangle and the best cost estimate pacts of the averted onsite costs were evaluated sepa-forming the peak of the triangle.The triangular distribu-rately.
tion is used in cost applications where only three points.
. best, and high-cost estimates, arc known (Ref. 23).
The averted onsite costs of Subsection 5.2.10 are sub-tracted from the costs listed in Tab!cs 4 and 6 resulting in The equations used in the uncertainty quantification are the net cost of performing the modifications. The cost-as follow benefit impacts of subtracting the onsite averted costs from each option are shown in Table 7.
Cost /Henefit Cost / delta R
=
delta R R
-R.
=
5.4 Cost-Ilenefit Uncerta. ties be i
in Where The uncertainty of the data has two effects on the cost-Change m. Risk or Risk i eduction benefit results.The first effect is on the value of the risk delta R
=
reduction; the second effect is on the value of the cost.
R Hase Case Risk
=
Hence, the uncertainties associated with the cost data and Ni = Option i Risk risk reduction data determine the uncertainties of the cost-benefit values.
The risk (R), base case or option, is defined as follows:
Uncertainty data were gathered, evaluated, and reported in the form of distributions for all data used in this analy-R = Core Damage Frequency x Consequences sis.This data-gathering and reduction is used to gauge the effects of the individual data uncertainty on the final cost-Risk reduction is defined as the change in risk from a base benefit modeling results of the analysis. There is also un-case model and the option models. The risk reduction un-certainty associated wit h the modeling in the analysis. It is certainty has two contributions that have two different ef-assumed that the model and modeling assumptions ade-fccts on the risk reduction uncertainty.These two effects Table 6. Cost. benefit summary without averted onsite costs
($/ person. rem reduction)
Person. Rem Reduction Person. Rem Reduction 3
Cost SK Without Operator Scram Cost. Benefit With Operator Scram Cost. Benefit (30 Plants)
(30 Plants)
(S/ Person. Rem)
(30 Plants)
($/ Person. Rem) i Option 1 1,500 3,189 470 193 7,750 1
Option 2 2,430 3,944 616 242 10,048 Option 3 3,960 4,700 843 4,036 981 Option 4 32,500 10,000 3,250 673 48,299 Option 5 6,030
-1,869
-b
-1,803
-b Optioti 6 7,290 6.504 1,120 4,112 1,77 nUsing costs from Table 4.
bNo tienefit; see text.
Table 7. Cost benefit including averted onsite costs ($/ person rem reduction)a Person. Rem Person. Rem Reduction Without Reduction With Cost SK Operator Scram Cost.Ilenefit Cost 51(b Operator Scram Costdienefit b
(30 Plants) (30 Plants)
($/ Person Rem) (30 Plants) (30 Plants)
($/ Person. Rem)
Option 1 1.020 3,189 320 1,470 193 7,616 Option 2 1,830 3,944 464 2,394 242 9,893 Option 3 3.240 4,700 689 3,360 4,036 833 Option 4 31,000 10,000 3,100 32,398 673 48,140 5,757
-1,803
-C Option 5 5,751
-1,869
-C Option 6 6,750 6,504 1,038 6,690 4,112 1,626 "The result of subtracting the averted onsite costs changes the emi-tienefit results, but it does not change their pminon relative to the 51.000/
person-rem nominal cat-benefit screemng value.
Using cost from Tables 4 amt 5.
No benefit; se e text.
l 1
l come from (1) those risk contributors that remain the all other risk model contributors represented by con-same from the base case to the options and (2) those risk stants).The results of the cost-benefit uncertainty analy-contributors that change from the base case to the op-sis are shown in Table 9.
tions. Those risk contributors that do not change from base case to the options only affect the uncertainty in the The resulis of the cost-benefit uncertainty analysis show absolute value of the total ri:,k due to an ATWS event.
that none of the means of the cost-benefit distributions l
The risk contributors that change from base case to the are below the $ 1,000 per person-rem guideline. Ilowever, options affect the uncertainty of the risk reduction (frac-without operator action, Option 1, for example, has a 26 tion of total risk reduced) from the base case. Therefore, percent probability of being at S I,000/ person-rem or less.
the two uncertainties in the risk analysis are those that A comparison of the uncertainty ranges associated with I
affect the absolute value (total) risk associated with the cach option's cost-benefit result is shorn in Figure 1.The ATWS sequence and those that affect the relative risk uncertainty distribution results reflect the uncertainty in (risk reduction) from base case to or' ions.
the data used to quantify the cost-benefit ratios. The un-certainty in the cost distributions and the consequence in all cases (base and options), the same model is used distributions tends to drive the mean values of the cost-with the only difference being the elimination or addition benefit uncertainty distributions to values higher than the of some cutsets of the reactor trip reliability failure fre-point estimate results.To understand why this occurs, the quency.
underlying cost and consequence distributions must be examined.
The uncertainty in the base case risk analysis is analy/cd to assess the uncertairsty bounds on the total risk due to The cost distribution is asymmetric with a hip,h probability the ATWS sequences. The event sequences and conse-density on the low cost side and a tail stretched out to the quence analysis distributions for the base case are run high cost estimate. The consequences distribution is also using Mond Carlo sampling to arrive at the results in Ta-asymmetric with a high probability density on the low con-ble 8. The distribution graphs (density and cumulative sequences side with some probability toward high conse-distributions) are shown in Appendix C of Reference 4.
quences. Iloth of these uncertainty distributions (cost and consequences)cause the Monte Carlo random simulation The risk reduction from tiac base case to tbc options is cal-to yield cost-benefit uncertainty mean values higher than culated by subtracting the option risk f rom the base case the point estimate values.
risk. Therefore, the uncertainty in risk reduction is deter-mined by contributors to reactor trip reliability in risk The uncertainties associated with the risks and costs of changes from the base case to the options, with all other this cost-benefit analysis are larpe. Therefore, the point contributors remaining constant. The cost-benefit uncer-estimate results have a wide uncertainty associated with tainty is evaluated with Monte Carlo sampling using the Ihem; hence, they have a potentially large error factor. A cost distributions and the risk reduction distribution (rep-comparison between the point estimate cost-benefit re-resented by those cutsets that changed in each option with suits and the uncertainty cost benelit results is shown in NURl:G-1341 10 L_
Table 8, Ur ase risk uncertainty results (person rem)
Mean 5th 50th 95th Without Operator Action 28,567 1,100 10.900 108.600 With Operator Action 16,831 600 5,700 59.100 Table 9, Cost benefit unecrtainty results Cost /llenefit ($/ Person. Rem Reduction)
Distribution Parameters
'7< Probability G Probability Option Mean 5th 50th 95th from $0 to $1,000 More than $1,000 Without Operator Action 1
6,750 201 2,280 26,200 26 74 2
10,020 278 3,320 38,300 20 80 3
22,100 600 6,010 85,600 8
92 4
42,200 1,300 14,300 167,000 4
96 5
No positive benefit 6
131,000 7,100 70,900 505,000 0
100 With Operator Action 1
146,000 4,430 44,000 579,000 0
100 2
159,000 4,600 50,600 649,000 0
100 3
7 8,0(10 3,400 33,900 301.000 1
99 4
526,000 17,300 17),000 2,030,000 0
100 1
i S
No positive benefit 6
127,000 4,100 41,400 467,000 0
100 Table 10. The uncertainty retribution mean values are bution would be on both sides of 7ero, both negative and higher than the point estimate mean values-positive. If the result was more positive than negative, 17or comoarison purposes, the cost-benefit uncertainty then the cost benefit of Option 3 would be less than that results of Options 6 and 3 were evaluated to assess which of Option 6 making Option 3 more favorable or vice versa option would be more favorable.The cost-benefit uncer.
for Option 6. As shown in Table 11, with eredit for opera-tainty of Option 3 was subtracted from the cost-benefit tot action, neither option is more favorabic than the uncertainty of Option 6. The resulting uncertainty distri-other.
1I NUlti!O-1341
il jiI
.I'o 1
6+
Xi 0
2
- o 5
.It
+
5 5
yi 5
+
+
5
+
t i
a 9
7i K.
+
S..
7 e
t 2
C 5
5 5
2 S
+
4 E,i 1
)
0 n
3 o
5 it 5
5 E
i 5
+
+
t 9
+
c
+
5 u
t
+
yi 3
7.
d 7l t.
s.
6 t
1 4
1 2
5 e
J 4
1
+
.3o 1
R
+
t c;
K i
t s
M s
f a
f E
4+
R 4
+
7l
+
s t
r; 2
+
n n
4 h
+
a 4
4 Xi o
o
!5 K;
E,;
7 s
+
i t
J t
1 r
u 9
6 e
u P
b 2
i
/
r t
is
.5'o
(
n; d
3 t
o t
3
+
i t
i y
+
t.i f
1t f.
3 3
3 e
n 7
7
+
+
+
n ia 6
Xi Mi 3
M.
e t
+
1 D
S r
D 4
N N
4 4
M.
e 4
/
c 4
E A
3 n
t I
s G
W.
o u
EL C
if t
3*
en Mi
.3'o e
b 1
t t
2 s
+
n o
n M
m m
C e
e e
e 1
h 2
c d
+
e I t 2
s 1
x.;
r
+
r u
t; e
=
g 2
0 iF 2
e C
c
.I'o P
1 NO TIC A
1 2
3 4
5 6
1 2
3 4
5 6
?
tO T
n n
n n
n n
n n
n n
n n
o o
o o
o io o
o e
o o
A o
it it it it it t
it it it it it w
itp p
p p
p p
p p
p p
p p
gO O
O O
O O
O O
O O
O O
(
TJRCH T
I W
2C%FOi5' O
Table 10, Point estimate / uncertainty results coruparison I
i Cost /11enefit ($/ person rem reduction)
Option Point Estimate Mean Distribution Mean Without Operator Action 1
320 6,750 2
464 10,200 3
989 22.100 4
3,10d 42,200 5
No benefit No benefit 6
1,038 131,000 With Operator Action 1
7,616 146,000 2
9,893 159,000 3
883 78,200 4
48,140 526,000 l-5 No benefit No benefit l'
6 1,626 127,000 1
1 Table 11. Option 6 versus Option 3 cost. benefit uncertainty comparison Cost ilenefit ($/ person rem)
Percent Fercent Mean 5th 50th 95th
<0
>0 Without Operator Action 99,300
-37,900 10,200 367,00(1 42 58 With Operator Action 46,200
-36,000
-718 184,000 50 50
- 6. DECISION RATIONALE provide a reduction in risk corresponding to a range of re-duction in CDP by 1.0E-7 :o 2.1E-6 events /Rx-year and a cost-benefit ratio generally higher than the This generic issue was identified as a consequence of the
$1,000/ person-rem nominal screening value for backfit 1983 Salem ATWS events and Generic lEter 83-28 considerations. A particular area of concern that was ex-(Ref. 2) to evaluate certain options for increasing the rcli-amiricd as part of this issue was the reliability of the ability of the reactor trip function in E plants. As part of RTTls. The RTI3s used in E plants are E-supplied break-this evaluation, she operational experience with two key ers of the DS-416 and DIl-50 scrhs, appropriately modi-components of the E SSPS received particular attention:
ficd for use in the reactor protection system. There has the UV driver cards and the RTIls. Six options to enhance been no appreciable change in the failure rates of these the reliability of the reactor trip function were evaluated breakers since the A1WS rulemaking analytis (Ref.1).
and are described in Section 3 of this report and in Refer-This matter is under active consideration by NRC as part ence 4. In the interest of completeness, a brief description of the requirements of items 4.2.3 and 4.2.4 of Generic is also given in the discussion below where the bases and Ietter 83-28 (Ref. 2) and is the subject of an ongoing rationale for their respective dispositions are presented.
evaluation of a proposed action regarding R'111 life test-With the exception of one option, which would result in ing and refurbishment or replacement. It is also under an increase in risk if implemented, all other options would evaluation as part of a larger research program on the ag-13 NUREG-1341
l l
1 ing assessment of E DS-Series circuit breakers used in all driver card. These relays would receive the same reactcr plant systems and not only the reactor protection system.
trip inputs from the SSPS logic cards and provide driver relays for the RTBs' UV trip and shunt trip devices (Refs.
Based on the relatively small predicted risk reduction 6 and 8), basically duplicating the function of the UV from the evaluated options and the results of the cost.
driver card by diverse means.
benefit analysis (see Table 10) no backfit requirements are justified in accordance with the backfit rufe,10 CFR As pointed out earlier, the safety improvement that may Part 50.109(a)(3).
result from implementing this option is very small (a CDF reduction of 2.0E-07/Rx-year)with an unfavorable mini-In the course of evaluating this generic issue, and the six mtim cost-benefit ratio of $9,900/ person-rem. Further-options in particular, we have gained certain insights that more, if the recommendations under Option 1 are imple-could be useful in improving the reliability and perform.
mented, as some E licensee; are now doing, the aced for ance of the reactor protection system.These insights may a diverse and redundant UV trip path would further di-be the subject of consideration by industry for possible minish. Hence, we conelude that this option should not be voluntary changes in design and procedurcs, including made a regulatory requirement.
technical specifications. If properly implemented, they could improve, or at least maintain an acceptable level of, 6.3 Option 3 - Overcurrent/ Fusible the reliabihty of the reactor protection system, reduce the regulatory burden on the licensecs, and extend the life of
- blUb, the RTBs.These insights are presented at the end of this This option entails the incorporation of design changes /
section following the discussion of each of the six options.
additions to provide a diverse tripping mechanism redun-These discussions include certain deterministic consid-dant and diverse to the RTBs that would create an over-erations germane to the disposition of each option.
current condition on receipt of a reactor trip signal, causing the opening of a fusible link placed in series with 6.1 Option 1 - New UV Driver Card the RTBs (Ref. 24)-
This option entails the replacement of the UV drivercard The risk reduction and cost-benefit quantification for this wi:h one that was modified and recommended by E (Ref.
option show that this option rated relatively better than 7). The new UV driver card has a fuse on the 48-volt out.
most of the other options evaluated, but it would still re-put and would have prevented four of the five short.
sult in a relatively small improvement in safety (a CDF re-circuit failures experienced by the original UV driver card duction of 2.lE-06/Rx-year) with an unfavorable cost-design.
benefit ratio estimated to be closer to the distribution mean value of $78.000/ person-rem than the point esti.
Even though the criteria for backfit requirements are not mate of $833/ person-rem.
met on the basis of the PRA results, we believe that the relatively low cost, the simplicity, and the effectiveness of The addition of a fusible link in series with the RTBs and this design change in eliminating the most troublesome the introduction of more components with poorly known failure modes of the UV driver cards make this W.
reliabilitics and associated costs would likely result in a recommended option worthy of serious considerationIy smaller safety benefit, and costs larger than those esti-E plant owners. A survey of the 30 E plants with the mated in Reference 19. An alternative approach based on SSPS design indicates that several plants have already re.
using the existing component may be far more cost-and placed or arc in the process of replacing the old UV driver safety-effective as discussed later under recornmenda-cards with the new design cards along with the g.
tions. We conclude that this option should not be made a recommended changes in test procedures (Ref. 7). We regulatory requirement.
endorse the E recommendations contained in Refer-ence 7. Ilowever, based on the results of our analyses, we 6.4 Option 4 - Relay Logic System conclude that this option cannot be made a regulatory requirement because of the associated small safety This option provides a diverse trip logic composed of a re-improvement (a CDF reduction of LUE-07/Rx-year)and lay logic system in parallel with the SSPS reactor trip logic an unfavorable minimum cost-benefit ratio of and respective UV driver cards. This dtverse logic system
$7,600/ person-rem.
would receive the same input from the analog instrumen-tation input relays as the SSPS logic and provide a redun-6.2 Option 2 - Diverse and Redundant New Relay UV Driver W naluation of this option confirms the ATWS rule provision exempting E plants from the requirement to This option provides a diverse and redundant UV trip provide a separate and independent reactor trip system path consisting of relays arranged in parallel with the UV such as that of this option.The results of the risk and cost-NURI:G-1341 14 l
1 i
l
benefit analyses (a CDF reduction of 4.9E-07 with a cost-Conclusions and Insights for Further Work benefit ratio of at $48,140/ person-rem) show that no backfit based on this option is warr:mted. Hence, we On the basis of the technical findings of the risk and cost-conclude Ihat this option should not be made a regulatory bencht analyses performed, and the discussion of each requirement.
opt,on presented in the preceding paragraphs, we con-clude that no backfit requirements arc warranted in ac-cordance with the backfit rule,10 CFR Part 50.109(a)(3).
6.5 Option 5 - Redundant Shunt Trip However, based on insights gained during the technical Coil na u tion of this issue in general, and the six options in particular, the following insights should be considered:
'Ihis option provides a redundant shunt trip configuration Decreasing the Irfil test frequency in conjunction for cach 10111 by replacing the UV trip mechanisms on
+
cach I' : 11 with a shunt trip mechanism (Ref. 9). The with the addition of an automatic trip function of the shunt trip de power supply would be st.pplemented with a contactors supplying the field current to the motor-fail-safe capacitor circuit that would supply a de charge to percrator (M/G) sets, and/or the M/G sets output operate the shunt trip on loss of the de power supply, breakers, These changes if implemented, would con-tribute toward (1) reducing the regulatory burden on The evaluation of this option indicated that itsimplemen.
the affected licensees and applicants, and (2) ex-tation would result in an increase in risk. This is duc,in tending the life of the RTIls, as well as providing a part, to the fact that the replacement of the UV trip coil diverse and redundant interruption of power to the mechanism with a shunt trip type of mechanism would re-control rods, thus improving, or at least maintaining, move the element of diversity between the UV and shunt the reliability of the reactor trip function.
trip coil mechanisms used in the present design of the irrlls. Ilence, we conclude that this option should not be A recommendation to study whether the RTil test-made a regulatory requirement.
ing frequency can be decreased has been made by NRR in its assessment of the feasibility of reducing technical specifications suncillance requirements 6.6 Option 6 - Contactors (Ref. 25).
Under this option, one RTH and its corresponding bypass The licensees proposing to adopt an approach such breaker would be replaced each by a contactor (Ref.10).
as the above should be allowed to do so assuming that the recommendations contained in ETechnical The risk reduction and cost-benefit quantification for this flulletin NSIDJHI-85-16 (Ref. 7) have been imple-option show that this option rated relatively better than mented. These recommendations, developed in ac-most of the other options evaluated, but it would still re-cordance with the requirements of 10 CFR Part 21, sult in a relatively small im provement in safety (a CD Fre-have already been implemen ted in several E plants, duction of 2.1E-06/Rx-year with a cost-benefit ratio esti-mated to be closer to the distribution mean value of Incorporation of the above insights in the design of
$127,000/ person. rem than the point estimate of the advanced IER (AlER) plant proposed by
$1,626/ person-rem). The replacement of one RTH and EpRI. Incorporation of these design features at this one bypass trip breaker in the present design would intro-early stage of the Al.WR design would be more effi-duce components of a different physical configuration cien'tly implemented than in a backfit setting.
with the distinct possibility of introducing an unquantified reduction in safety improw ment associated with the physical and functional integrity si their interface with
- 7. IMPLEMENTATION the existing system components and configuration. We conclude that this option should not be made a regulatory No actica imposing new regulatory requirements is nec-requirement.
essary for resolution of this issue. A distribution of Isefer-ence 4 and the regulatory analysis NUREG report is As discussed under Option 3, an ahernative apprc,ach n ade to include all E licensees. The insights contained in based on using existing components may be far mc..e cost-Section 6 could form the basis of industry initiatives and
.and safety-effcctive. This alternative approach is dis-subsequent discussions between NRC and the affected li-cussed below under conclusions and recommendations.
censees and applicants.
l l
l 15 NUREG-1341
i REFERENCES i
1.
FederalRegister. Vol. 49, No.129, pp. 26036-26045,
- 12. USNRC, " Amendments to 10 CFR 50 Related to "10 CFR Part 50, Reduction of Rish from Antici-Anticipated Transients Without Scram (ATWS) pated Transients Without Scram (ATWS) Events Events," SECY-83-293, July 19,1983.
for Light. Water Cooled Nuclear Power Plants,"
June 26,1984.
- 13. Richard P. Ilurke et al., " Economic Risks of Nuclear Power Reactor Accidents," NUREG/CR-3673, 2.
NRC Letter to All Licensecs of Operating Reactors, S AND84-0178, May 1984.
Applicants for Operating Licenses, and Holders of Construction Permits," Required Actions Ilased on
Accident Risk in U. S. Commercial Nuclear Power Generic Letter No. 83-28, July 8,1983.
Plants," WASil-1400 (NUREG-75/014), October 1975.
3.
Memorandum from R. Ilcrnero to T. Speis, "En-hancement of the Reliability of Westinghouse Solid
- 15. J. W. Ilickman,"PRA Procedures Guide. A Guide State Protection Systems (SSPS)," April 5,1985.
to the Performance of Probabilistic Risk Assess-ments for Nuclear Power Plants," American Nu-4 D. A. Reny et al.," Evaluation of Generic Issue 115, clear Society and Institute of Electrical and Elec-linhancement of the Reliability of Westinghouse tronics Engineers, NURl!G/CR-2300 (2 of 2),
Solid State Protection System," NUREG/CR-5197, January 1983.
EGG-2546, January 1989.
16.
A. S. Benjamin et al., " Evaluation of Severe Acci-dent Risks and the Potential for Risk Reduction:
5.
W. II. Andrews et al., " Guidelines for Nuclear Surry Power Station, Unit 1," Vol. It Sequoyah Power Plant Safety issuc Prioritization Information P wer Station, Unit 1, Vol. 2, Draft NUREG/
L)evelopment " NUREG/CR-2800, PNL-4297, CR-4551. February 1987.
Supplement 1, May 1983.
- 17. J. R. Ball et al.,"A llandbook for Cost Estimating:
6.
Memorandum from 11. R. Denton to 'I,.1,. Spets, A Method for Developing Estimates of Costs for
" Schedule for Resolving and C ompleting Genenc Generic Actions for Nuclear Power Plants,"
Westinghouse Solid State Protection System,,,{
issue No.115, 'linhancement of the Reliability o NUREG/CR-3971, ANilEES-TM-265. October 3934*
July 7,1986.
- 18. Science and Engineering Associates, Inc., et al.,
7.
Westinghouse 'I.cchnical llulletm, "SSPS Under-
" Generic Cost Estimates: Abstracts from Gencric voltage Output Driver Card," NSID-lW85-16, Studies for Use in Preparing Regulatory impact July 31,1985.
Analyses " NUREG/CR-4627, June 1986.
8.
U.S. Nuclear Regulatory Commission (NRC), "A
- 19. Memorimdum from C. Morris to D. L. Hasdekas, Prioritization of Generic Safety issues,"
" Estimated Cost of Option 3 Enhancement of the NUREG-0933, Supplement 6, pp. 3.115-1 through Reliability of Westinghouse SSPS " August 31, 3.115-7, March 1987.
1988.
9.
Memorandum from F. Rosa to K. Kniel,"Reliabil-20.
D. P. Mackowiak et al.," Development of Transient ity of W SSPS," March 30,1988.
Initiating Event Frequencies for Use in Probabil-istic Risk Assessments," NUREG/CR-3862,
- 10. Memorandum from F. Rosa to D. L. Hasdekas.
EGG-2323, May 1985.
" Reliability of Westinghouse Solid State logic Protection System: Contractor Draft Report on GI
- 21. Memorandum from A. J. Dipalo lo G. R. Mazetis, 115," August 25,1988.
on Cost Estimates for Generic issue 125.11.7, dated February 5,1988.
I1. Kenneth D. Russell and Martin H. Sattison,"Inte-grated Reliability and Risk Analysis System
- 22. S. W. Heaberlin et al., "A llandbook for Value-(IRRAS) Version 2.0 Users Guide " NUREG/
Impact Assessment,"
N U REG /CR-3568, CR-5111 May 1989 PNI -4646, December 1983.
NURiiG-1341 16
- 23. @ RISK," Risk Analysis and Modeling for the PC,"
- 25. Memorandum from C. E. Rossi to S. Varga et al.,
Palisades Corporation, March 1,1988.
" Review of Report on the Feasibility of Reducing Technical Specifications Surveillance Testing,"
- 24. Memorandum from C. Morris to F. Rosa," Westing-N vember 21,1988.
house Reactor Trip Breakers (RTB) Diversity,.
November 23,1988.
)
I l
l 1
i
)
I l
I 1
l j
1 17 NUREG-1341 i
i
NRC FORM 335 U.S.WUCLE AR REGULATORY COMMISSION
- 1. REPORT NUMBE R Mc 11C2.
NnYum N m rhin,[
- w. 22m BIBLIOGRAPHIC DATA SHEET (See instructions ort the reverse)
NllREG-1341 2.Tn LE AND SUBTITLE Regulatory Analysis for the Resolution of Generic Issue 115, Enhancement of the Reliability of the Westinghouse
^["'""'*'S"[
Solid' State Protection System May 1989
- 4. F IN OR GR ANT NUMBE R b AUTHOR (S) 6 TYPE OF REPORT Demetrios L. Basdekas Technical 7 VL R ROD COV E RL D Hmiume 0.rrea 8
RF RMING OR NfZ AT ION - NAML AND ADDRL 55 tir NRC. orove Dwnston. orroce or Regoon. us Navnar negatorary commoss,on, ene mornns atur,eu ir rontextor. provisi.
Pivision of Safety Issue Resolution I
Office of Nuclear Regulatory Research lJ.S. Nuclear Regulatory Commission Washington, D.C.
20555
- 9. SPONSOR 1NG ORG ANIZATION - NAME AND ADDRLSS tir knc. rype "same m etxwe, ucentrator.orove Nac owa on. orrece ar ney,an. us Noctear neautorary commouron.
and man,ny nodous.1 Division of Safety Issue Resolution Office of Nuclear Regulatory Research
'l.S. Nuclear Regulatory Commission Washington, D.C.
20555
- 10. SUPPLEMENTARY NOTES l
- 11. ABST 9 ACT (100 worm or eass Generic Issue 115 addresses a concern related to the reliability of the Westinghouse reactor protection system for plants using the Westinghouse Solid State Protection System (%PS). Several options for improving the reliability of the Westinghouse reactor trip function for these plants and their effect on core damage frequency (CDF) and overall risk were evaluated.
This regulatory analysis includes a quantitative assessment of the costs and benefits associated with the various options for enhancing the reliability of the Westinghouse SSPS and provides recommendations for consideration as industry initiatives. No new regulatory requirements are proposed.
u x e v wO R os o t Sc R m i O s s u,,, wo,o, o,,,,,,,e,n, -u.o,,,
~~,c,,,,,,n <w,,,n, une,eo,,,,,
u ~, -.n si m w
- i Reactor Protection System, Reactor Trip Breakers, Undervoltage
.,,ds..,,,,
Driver Card, Anticipated Transients Without Scram l
l linclassified i
i r,,m,mr,:
1 linclassified 1
t b. NUMBL R Of FAGE S 16 PhlCL l
I NRC FORV 33512491 i
l f
1
~
UNITED STATES -
1
! NUCt. EAR REGULATORY COMMISSION -
'"'E/Ocf0EsNII' l WASHINGTON, D.C. 20555 :
PERMlf No G E7 p-OFFICIAL BUSINESS
[ PENALTY FOR PRIVATE USE. $300 f11PGIEb 15 05551]o A.
gi/Qg.bpggt,1C AT 106 ate 2oss5 oc TP-S
(~g,gNGTON j
, v:
n; 31 I.4k 1h i
E!
OJ mi H;
)>j j
El 01Y si s1
-21 m!
<J
~ l N:
s f
c
_ _ - - _ _ _ _ - - _ _ - - - -