ML20148S600
| ML20148S600 | |
| Person / Time | |
|---|---|
| Site: | Calvert Cliffs  |
| Issue date: | 05/08/1997 |
| From: | BALTIMORE GAS & ELECTRIC CO. |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| Shared Package | |
| ML20148S368 | List:
|
| References | |
| RTR-NUREG-1560 NUDOCS 9707080269 | |
| Download: ML20148S600 (10) | |
Text
,
CHARLES H. CausE Baltimore Gas and Electric Company Vice President Calven Cliffs Nuclear Power Plant Nuclest Energy 1650 Calvert Cliffs Parkway Lusby. Maryland 20657 410 495-4455 May 8,1997 U. S. Nuclear Regulatory Commission Washington,DC 20555 ATTENTION:
Office of Nuclear Reactor Research
SUBJECT:
Calvert Cliffs Nuclear Power Plant Unit Nos.1 & 2; Docket Nos. 50-317 & 50-318 Comments on Draft NUREG-1560 " Individual Plant Examination Program:
Perstrectives on Reactor Safety and Plant Performance"
REFERENCE:
(a)
Letter from Mr. C. H. Cruse (BGE) to NRC Office of Nuclear Reactor Research, dated March 27,1997, Individual Plant Examination Program:
Perspectives on Reactor Safety and Plant Performance In Reference (a) above, Baltimere Gas and Electric Company submitted preliminary comments on the above referenced NUREG. We also stated our plans to submit additional comments prior to the final due date. The additional comments are provided in Attachment (1). A summary of our major concerns is provided below.
A large portion of this NUREG is dedicated to the comparison of the Individual Plant Examination numerical results between plants. Drawing conclusions without researching the design characteristics, assumptions, and quality of the Individual Plant Examinations beirg compared, could result in significant misinterpretation.
By publishing a document with conclusions which appear to represent comparisons of operating risk, the reader can easily be mislead. We therefore strongly recommend that comparison between plants be completely eliminated from NUREG 1560.
9707000269 970516 PDR NUREO
' b'"' ' "1560 C PDR TT/
Office ofNuclear Reactor Research t
May 8,1997 Page 2 7
Should you have questions regarding this matter, we will be pleased to discuss them with you.
)
i Very truly yours, h
/
l CHC/SJR/ dim
Attachment:
(1)
Comments on Draft NUREG-1560 " Individual Plant Examination Program:
Perspectives on Reactor Safety and Plant Performance" cc:
Document Control Desk,NRC H. J. Miller, NRC R. S. Fleishman, Esquire Resident Inspector, NRC J. E. Silberg, Esquire R. I. McLean, DNR Director, Project Directorate I-1, NRC J. H. Walter, PSC A. W. Dromerick, NRC I
s a
l i
l
+
1
j ATTACHMENT (1) l t
i i
1 i
4 i
j Baltimore Gas and Electric Company Comments on Draft NUREG-1560," Individual Plant Examination Program:
s' j
Perspectives on Reactor Safety and Plant Performance" A
i i
i i
i i
i.
l 4
1 5
l' i
I t
i Calvert Cliffs Nuclear Power Plant
~
Units 1 & 2 Msy 8,1997
1 ATTACHMENT t1) i i
BALTIMORE GAS AND ELECTRIC COMPANY COMMENTS ON DRAFT NUREG-1560 " INDIVIDUAL PLANT EXAMINATION PROGRAM:
l PERSPECTIVES ON REACTOR SAFETY AND PLANT PERFORMANCE" i
i I.
Section 1.3, page 1-3, paragraph 3 states: "In examining the IPEs Individual Plant Examination and developing this report, the stafused the information as reported by the licensees. That is, the i
staf did not consider the quality (e.g., occuracy) of the analyses when determining the i
i implications of the collective JPE results. Therefore, the staf used information from each IPE, even if a licensee's IPE/PRA was unacceptable (in part or overall), and no adjustment or modsfcation was made." Page 1-3 paragraph 7 states: "The IPE Insights Program is based solely on licensee submittals, which summarise the IPE analyses and do notfully document all design characteristics, analysis assumptions, and results. This limits the ability tofully account for similarities anddiferences in results. "
1 i
Baltimore Gas and Flectric Comnany mGE) Comment A large portion of this NUREG is dedicated to the comparison of the Individual Plant Examination t
(IPE) numerical results. Drawing conclusions without researching the design characteristics, t
assumptions, and quality of the individual IPEs being compared, may result in significant misinterpretation. Generic Letter 88-20 requested each licensee to "... perform a systematic examination to ider.tify any plant-specific vulnerabilities to severe accidents and report the results j
to the Commissica." NUREG-1560 should focus on the quality of systematic examinations rather than numerical results.
t
'Ihe Combustion Engineering Owners Group (CEOG) Probabilistic Safety Assessment (PSA)
Working Group (PSAWG) has found that the differences between the Combustion Engineering l
(CE) IPEs are due to many reasons, each requiring careful evaluation to determine whether the differences are due to modeling techniques or plant differences. When the differences are due to i
modeling, considerable effort is required to determine the most appropriate solutions in order to confirm the models.
By publishing a document with conclusions that appear to represent comparisons of operating risk, the reader can easily be mislead. The statement, "Calvert Cliffs 1 & 2 have a CDF well above other CE plants, primarily because of a higher dependence on HVAC [ Heating, Ventilation and Air ConditioningJ and less capability (relative to other CE plants) to remove decay heat through the steam generators or feed-and-bleed cooling," implies an in-depth analysis between the CE plants, which is not true. See Comment 2 below.
2.
Page 3-47,frstparagraph, states: "Calvert Chfs 1 & 2 have a CDF well above other CEplants, primarily because of a higher dependence on HVAC and less capability (relative to other CE plants) to remove decay heat through the steam generator orfeed-and-bleed cooling. "
BGE Comment This conclusion agipears to overly simplify the characterization of Calvert Cliffs' higher CDF. A significantly more detailed analysis is needed to determine the real reasons for the differences in the calculated CDF, including a review of many of the issues identified in Section 6.2, j
Characteristics of a Quality Probabilistic Risk Assessment.
l 1
_~ -
_.7-__.--
- -..., _ - -. ~. -.
ATTACHMENT (1) i BALTIMORE GAS AND ELECTRIC COMPANY COMMENTS ON DRAFT NUREG-1560 " INDIVIDUAL PLANT EXAMINATION PROGRAM
- l-PERSPECTIVES ON REACTOR SAFETY AND PLANT PERFORMANCE" 5
t i
Reasons for a higher CDF could include:
Calvert Cliffs' comprehensive analysis ofinitiating events, including many support systems;
]
l j
Realistic and sometimes conservative success criteria; He inclusion of system dependencies, such as HVAC, service water and component cooling water make-up systems; i
{
Detailed system analyses which discovered many subtle dependencies; i
Extensive data collection which often resulted in higher failure rates and unavailability than t
generic data indicated; Plant-specific common cause analysis which often resulted in higher common cause factors than generic data; and A more rigorous flood analysis (see NUREG-1560, page 3-53).
Rese reasons have more to do with the quality of the analyses than the characteristics of the plant.
His higher CDF is likely due to a complex combination of thoroughness (quality), conservatisms, and plant configuration. He more thorough the analysis, the more issues that are addressed and, therefore, the more likely that these issues contribute to core damage. And although BGE intends Calvert Cliffs' Probabilistic Risk Assessment (PRA) to be as realistic as possible, conservatisms were included. These conservatisms result from both the desire to build a defensible and thorough analysis, and the desire to complete such analysis in a cost-and time-effective manner. We also believe it prudent to use bounding analysis and/or bounding modeling techniques for issues which would otherwise result in excessive resources to resolve. Finally, plant configuration, including its design, operation, and maintenance, also plays an important role. It is, therefore, impossible to draw a conclusion about Calvert Cliffs' risk relative to other plants without first resolving the issue of consistency between the analyses.
j The CEOG PSAWG is performing detailed reviews of the differences between the CE plant PSAs to determine the reasons for the differences between the CE plants' calculated CDFs. His process will require time to determine the reasons for these differences.
Baltimore Gas and Electric Company recommends that NUREG-1560 clearly characterize any comparison with the uncensinty of the consistency of approaches between the PSAs and the need to resolve such differences in a thorough and comprehensive manner. It is also important to note that the IPE submittals are a snapshot in time. Calvert Cliffs' PRA continues to improve and change as knowledge and understanding of the plant and issues are better understood.
i I
2
~ ~
' ~ ~
~
~
~
~ ~~
ATTACHMENT (1)
BALTIMORE GAS AND ELECTRIC COMPANY COMMENTS ON DRAFT NUREG-1560 " INDIVIDUAL PLANT EXAMINATION PROGRAM:
PERSPECTIVES ON REACTOR SAFETY AND PLANT PERFORMANCE" A more appropriate statement would be:
i Calvert Cl;ffs' higher calculated CDF is difficult to explain since its design is very similar to at j
least two other CE plants. Its model included a higher dependency on HVAC, which may be due to its treatment of this issue. The stated lesser capability (relative to other CE plants) to remove decay heat through the steam generators or feed-and-bleed cooling is not readily apparent and requires additional evaluation to resolve.
Specific discussion on Calvert Cliffs' HVAC and decay heat removal is provided below.
HVAC Section 6.3, page 6-13, states:
"A common example of a support system initiator that is not rigorously analyzed is the loss of Control Room heating, ventilation and air conditioning (HVAC) system." It also states: "However, the basis for eliminating some dependencies (primarily HVAC dependencies), in some cases, was qualitative and made without apparent calculation support."
In Calvert Cliffs' PRA, the HVAC systems were realistically modeled to provide the best understanding of the impact of the loss of Control Room and Cable Spreading Room HVAC, Switchgear Room HVAC, and Emergency Core Cooling System Room Cooling. The mere inclusion of these dependencies adds to accident sequences that contribute to CDF. This is an area where completeness might be the governing reason for the differences between Calvert Cliffs and other CE plants. In addition, ventilation room heat-up analyses are typically conservative. In the case of Calvert Cliffs Nuclear Power Plant's Switchgear Room HVAC, the 480 VAC and 4 kV busses were assumed failed when the t.ction to recover ventilation failed.
l Decar Heat Removal Calven Cliffs' Auxiliary Feedwater System (AFW) is comparable to that of other CE plants, as can be seen in Table 11.22. Its AFW System consists of two turbine-driven AFW pumps and one motor-driven AFW pump per unit. The other unit's AFW motor-driven pump may be cross-connected, as well. It is, therefore, not likely that a significant design or operational difference exists between Calvert Cliffs and other CE plants. However, Calvert Cliffs' PRA does address common cause failure between the pumps of the turbine-driven pumps and motor-driven pumps, and between the Unit I and Unit 2 pumps. Although it is uncertain as to whether this is a driving issue,it is likely a contributor.
It is true that Calvert Cliffs' power-operated relief valves (PORVs) are small, therefore limiting the time available to initiate feed-and-bleed for effective heat removal. However, some CE plants i
also have small PQRVs or no PORVs. Again, it is not believed that this represents a reason for the
{
difference in calculated CDF between Calvert Cliffs and other CE plants.
3
ATTACHMENT (1)
BALTIMORE GAS AND ELECTRIC COMPANY COMMENTS ON DRAFr NUREG-1560 " INDIVIDUAL PLANT EXAMINATION PROGRAM:
PERSPECTIVES ON REACTOR SAFETY AND PLANT PERFORMANCE" 3.
Page 3-50,frst bullet, states: "For example, the Calvert Chfs units require both PORVs for successfulfeed-and-bleed (their PORVs are small). As a result, these units do not always credit feed-and-bleed as a possible heat removal system in response to a loss of main feedwater event because the RCS [ Reactor Coolant System] will remain above the HPI[high pressure injection]
shutofhead unless the reactor is tripped within 10 minutes. "
BGE Comment ne statement of tripping the reactor in 10 minutes is not true. At the time of the IPE submittal, we estimated that there was approximately 10 minutes from the loss of all feedwater (and its associated reactor trip) to the time the steam generators reach -350 inches. This assumes feedwater is lost at the time of the trip. Reaching -350 inches is the criteria the operators use to initiate feed-and-bleed. He time frame from initiating event to -350 inches was believed to be too short to allow the initiation of feed-and-bleed. The updated Calvert Cliffs PRA shows a small probability of success for this scenario ofinitiating feed-and-bleed, before reaching -350 inches.
4.
Page 3-50, second bullet, states: "Calvert Chfs does not credit steam generator depressurisation and use cf condensate for heat removal because of the small single ADY [ atmospheric dump valve] fur each steam generator. "
BGE Comment j.
He updated Calvert Cliffs PRA now credits feeding the steam generators with the condensate system. The amount of improvement provided by the additional functionality is difficult to 1
estimate due to the significant number of changes made in the model between the IPE submittal and the current update.
5.
Page 3-51, bottom bullet, states: "Confguration diferences rangefrom units withjust one diesel per unit with a swing diesel between units (such as Calvert Chfs) to two diesels per unit with additional backup capability existing or being added (such as Palo Verde). Calvert Chfs is in the process ofadding diesels to improve their onsitepower reliability. "
BGE Comment Note that the IPE submittal did not credit the two new diesels. See Comment 9.
4
ATTACHMENT (la BALTIMORE GAS AND ELECTRIC COMPANY COMMENTS ON DRArr NUREG-1560 " INDIVIDUAL PLANT EXAMINATION PROGRAM:
PERSPECTIVES ON REACTOR SAFETY AND PLANT PERFORMANCE" 6.
Page 3-52, pfth paragrqph, states: "ATWS is a relatively low contributor to CDF (less than SE-6/ry)for all but one plant in this group (Calvert Chfs). In that single exception, the higher ATWS contribution rejects the analyses assumptions and the assessment that the plant has an unfavorable moderator temperature coeficientfor a largefraction of the time (40%). The result is an analysis that does not credit the mitigatingfeatures mentioned above, assuming instead that an ATWS leads directly to core damage. This pessimistic stance leads to an ATWS CDF that is a factor ofsix greater than the next highestplant ATWS CDFamong the CEplants. "
BGE Comment Anticipated transient without scram (ATWS) recovery with a favorable moderator temperature coefficient (MTC) has been added to the model. We have also further evaluated the impact of an unfavorable MTC. Calvert Cliffs' updated PRA now assumes an unfavorable MTC 22% of the time when condenser vacuum is available, and 26% when it is not. This higher value is awumed
. due to the loss of turbine bypass heat removal, and the higher steam generator pressure due to the dependence on the steam generator safety valves.
Calvert Cliffs' ATWS contribution is now consistent _with other CEOG plants.
7.
Page 3 33, prst paragraph, states: "Calvert Chfs performed a more rigorous analysis that accumulated the combined efects ofnearly a dozenfoodscenarios (rather than using a series of individual screening arguments) that collectively contribute to an internalfood CDF of about 1.5E-5/ry (a 6% contribution to total CDF). It is unclear whether thefooding contribution will be signifcantly higherfor other plants sf they use the screening approach employed by Calvert Chfs. "
BGE Comment This is an example of where the quality of the analysis appears to impact the bottom line number.
8.
Page 9-27, Section 9.3.2.2, frst paragraph, states: "Only Calvert Chfs identified a plant improvement in this area (RCP Seal LOCA). (Page 9-27). The implemented improvements in the CCWsystem shouldreduce thefrequency ofan RCP sealLOCA.
pGE,Cpmment Reactor Coolant Pump (RCP) Seal Loss-Of-Coolant Accident (LOCA) importance has been significantly reduced in the updated Calvert Cliffs PRA model. This is due to more realistic modeling of the high pressure safety injection pumps' dependency on component cooling water.
For RCP Seal LOCA break sizes, thermal-hydraulic analysis indicates that the recirculated Reactor Coolant System temperature will not exceed the high pressure safety injection pump seal temperature (the most limiting constraint) for which component cooling water is required.
Therefore, the dominant IPE sequence of loss of component cooling water challenging both the RCP seals and the high pressure safety injection pumps is no longer applicable.
5 L
ATTACHMENT (1) l BALTIMORE GA5 AND ELECTRIC COMPANY COMMENTS ON DRAFT NUREG-1560 " INDIVIDUAL PLANT EXAMINATION PROGRAM:
PERSPECTIVES ON REACTOR SAFETY AND PLANT PERFORMANCE" 9.
Page 11-81, lastparagraph, states: "Theformer diesel confguration at Calvert Chfs 1 & 2 was one dieseldedicatedper unit with one swing diesel betweenplants. The enhanceddesign credited in the submittalincluded the addition oftwo diesels at the site, onefor eachpha,;
1 BGE Comment Calvert Cliffs' IPE submittal did not credit the new diesels in its results. A sensitivity study, s
Calvert Cliffs Nuclear Power Plant IPE Summary Report, page 3.4.1-33, was included which estimated the reduction of the overall CDF at 17.7%. The statement above implies the IPE results 3
include this addition.
10.
Page 11-83, Section 11.3.2.5, secondparagraph, states: " Examination ofthe Calvert Clifs 1 & 2 i
submittal considers thefraction oftime that the reactor cores at these units have an unfavorable moderator temperature coe.[ficient (AffC), discussed below, to be about 40%, significantly above that normally seen in PWR [ pressurized water reactor] PMs. Whether this is a true design diference or afunction ofpessimistic analysis, can not be determinedjustfrom the submittal.
Partially as a result of this consideration, the Calvert Chfs 1 & 2 analyses did not model the mitigationfeatures mentioned above at all. This submittal assumes thatfailure to scram leads to core damage, a pessimistic assumption. Even with this bounding approach, ATWS barely contributes 10% to the totalplant CDFsfor bothplants.
l BGE Comment See Commcat 6.
11.
Chapter 14, Attributes of a Quality PM, page 14-2, prst paragraph states: "This chapter is not intended to prescribe guidelinesfor how to perform a quality PM. This chapter is only intended to provide the attributes of a quality PM such that a reader canjudge of a specific PM (e.g.,
JPEs) is a qualityPM.
BGE Comment
. Since Chapter 14 is not meant to be used to establish IPE quality requirements nor for establishing the quality requirements for future PRA applications, BGE recommends deleting it from NUREG-1560.
Chapter 15, Comparison of Individual Plant Examinations To A Quality Probabilistic Risk Assessment, uses Chapter 14's attributes of a quality PRA. Since these attributes do not represent those which were required to meet the intent of Generic Letter 88-20, BGE also recommends deleting this chapter.
6
ATTACHMENT (1)
BALTIMORE GAS AND ELECTRIC COMPANY COMMENTS ON DRAFT NUREG-1560 " INDIVIDUAL PLANT EXAMINATION PROGRAM:
PERSPECTIVES ON REACTOR SAFETY AND PLANT PERFORMANCE" 12.
Chapter 16, Safety Goal Implications, first paragraph states: Chapter 16 provides a more in-depth discussimt including the approach adopted to infer how the IPE results were compared to the quantitative health objectives.
BGE Comment In SECY-90-104," Role ofIndividual Plant Examinations (IPE)In Assessing Industry Status with Respect to the Commission's Safety Goal Policy," dated March 20,1990, the NRC staff state 6 L
" Based on the significant additional resources that would be required to make a meaningful comparison of the IPE results with the Safety Gal Policy Statement and the potential problems associated with using the as-submitted IPE data, the staff recommends that no direct ccinparisons be made unless the IPEs are reviewed to a greater level of detail than currently planned. Therefore, no guidance or criteria for such a comparison are being proposed at this time. This recommendation is consistent with original purpose of the IPE which was to ensure a systematic evaluation of each plant for vulnerabilities to severe accidents. Enclosure 2 provides additional background on the purpose of the IPE as it has been communicated to licensees. However, if the Commission desires to invest the resources required for such a review, the staff will prepare a plan, including guidance and criteria, for Commission review and an assessment of the impact on other planned work."
It is BGE's understanding that a review of greater detail did not occur. Therefore, BGE recommends that the direct comparison of IPE's performed in Chapter 16 be removed from NUREG-1560.
7