ML032380047

Results of the Clinton Power Station Unit 1 SDP Phase 2 Notebook Benchmarking Visit

ML032380047
Person / Time
Site:	Clinton
Issue date:	08/25/2003
From:	Reinhart F Division of Systems Safety and Analysis
To:	O'Reilly P, Richards S NRC/NRR/DIPM, NRC/RES/DRAA
Wilson P, NRC/NRR/DSSA/SPSB, 415-1114
References
Download: ML032380047 (20)

Text

August 25, 2003 MEMORANDUM TO: Stuart Richards, Chief Inspection Program Branch Division of Inspection Program Management Office of Nuclear Reactor Regulation Patrick D. OReilly Operating Experience Risk Applications Branch Division of Risk Analysis and Applications Office of Nuclear Regulatory Research FROM: Mark F. Reinhart, Section Chief /RA/

Licensing Section Probabilistic Safety Assessment Branch Division of Systems Safety and Analysis Office of Nuclear Reactor Regulation

SUBJECT:

RESULTS OF THE CLINTON POWER STATION UNIT 1 SDP PHASE 2 NOTEBOOK BENCHMARKING VISIT During February, 2003, NRC staff and contractors visited the Exelon Generation Company office in Warrenville, IL to compare the Clinton Significance Determination Process (SDP)

Phase 2 notebook and licensees risk model results to ensure that the SDP notebook was generally conservative. The Clinton PRA did not include most external initiating events (only modeled fire initiators); so no sensitivity studies were performed to assess the impact of these initiators on SDP color determinations. In addition, the results from analyses using the NRCs draft Revision 3i Standard Plant Analysis Risk (SPAR) model for Clinton were compared with the licensees risk model. The results of the SPAR model benchmarking effort will be documented in next revision of the SPAR (revision 3) model documentation.

The benchmarking visit identified that there was good correlation between the Phase 2 SDP Notebook and the licensees PRA. The results indicate that the Clinton Phase 2 notebook was generally more conservative in comparison to the licensees PRA. The revision 1 SDP notebook will capture 88% (results matched or overestimated the licensees PRA by one order of magnitude) of the risk significance of inspection findings. A summary of the results of comparisons of hypothetical inspection findings between the SDP notebook and the licensees PRA are as follows.

Attachments: As stated CONTACT: Peter Wilson, SPSB/DSSA/NRR 301-415-1114

S. Richards 2 P. OReilly 5% Underestimates Risk Significance 72% Match Risk Significance 16% Overestimates Risk Significance by 1 Order of Magnitude 7% Overestimates Risk Significance by 2 Orders of Magnitude The Rev-1 SDP notebook has been significantly improved as a result of the benchmarking activity. The number of cases that the Rev-1 SDP would match that of the updated licensees PRA has increased from 35 to 72 percent. In addition, the number of underestimations decreased from 35 to 5 percent. However, the number of overestimations slightly increase from 14 to 23 percent.

The licensees PRA staff was very knowledgeable of the plant model and provided very helpful comments during the benchmark visit.

Attachment A describes the process and results of the comparison of the Clinton SDP Phase 2 Notebook and the licensees PRA.

Attachments: As stated

S. Richards 2 P. OReilly 5% Underestimates Risk Significance 72% Match Risk Significance 16% Overestimates Risk Significance by 1 Order of Magnitude 7% Overestimates Risk Significance by 2 Orders of Magnitude The Rev-1 SDP notebook has been significantly improved as a result of the benchmarking activity. The number of cases that the Rev-1 SDP would match that of the updated licensees PRA has increased from 35 to 72 percent. In addition, the number of underestimations decreased from 35 to 5 percent. However, the number of overestimations slightly increase from 14 to 23 percent.

The licensees PRA staff was very knowledgeable of the plant model and provided very helpful comments during the benchmark visit.

Attachment A describes the process and results of the comparison of the Clinton SDP Phase 2 Notebook and the licensees PRA.

Attachments: As stated CONTACT: Peter Wilson, SPSB/DSSA/NRR 301-415-1114 Distribution: SPSB: r/f Accession#ML032380047 G:\spsb\wilson\clintonbenchnrc.wpd NRR-096 OFFICE SPSB SPSB:RIII SPSB:SC NAME PWilson:nxh2 SBurgess MReinhart DATE 08/20/03 08/21/03 08/25/03 OFFICIAL RECORD COPY

SUMMARY

REPORT ON BENCHMARKING TRIP TO THE CLINTON POWER STATION UNIT 1 (Feb. 24 - 28, 2003)

M. A. Azarm Energy Sciences and Technology Department Brookhaven National Laboratory Upton, NY 11973-5000 ATTACHMENT

Table Of Contents Page

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2. Summary Results from Benchmarking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

3. Additional Proposed Modifications to SDP Worksheets . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.1 Specific Changes to the Rev. 0 SDP Worksheets for the Clinton SDP . . . . . . . . . 11 3.2 Generic Changes in IMC 0609 for Guidance to NRC Inspectors . . . . . . . . . . . . . . 13 3.3 Generic Change to the SDP Notebooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

4. Discussion on External Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

5. List of Participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 List of Tables Page Table 1: Summary of Benchmarking Results for Clinton . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Table 2: Comparative Summary of the Benchmarking Results . . . . . . . . . . . . . . . . . . . . . . 10

-ii-

1. INTRODUCTION This report compares the NRC Risk Informed Inspection Notebook, developed by BNL, and the licensees risk model for the Clinton Power Station to ensure that the Notebookwas generally conservative. A benchmarking of the Rev. 0 SDP Notebook for the Clinton NPP was conducted between February 24 to 27, 2003 at the Exelon Generating Company office in Warrenville IL.

NRC staff (M. Parker, M. Reinhart, and P. Wilson) and BNL staff (A. Azarm ) participated in this Benchmarking exercise.

In preparation for the meeting, BNL staff reviewed the SDP notebook for the Clinton Power Station and evaluated a set of hypothetical inspection findings using the Rev. 0 SDP worksheets. In addition, NRC staff provided the licensee with a copy of the meeting protocol.

The major milestones achieved during this meeting were as follows:

1. Recent modifications made to the Clinton PRA were discussed for consideration in the Rev. 1 model to be prepared following benchmarking.

2. Importance measures, including the Risk Achievement Worths (RAWs) for the basic events in the internal event model for average maintenance, were obtained from the licensee.

3. Benchmarking was conducted using the Rev. 0 SDP model and the revised SDP model considering the licensees input and other modifications that were judged necessary based on comparison of the SDP model and the licensees detailed model.

4. For cases where the color evaluated by the SDP notebook differed from that determined based on the RAW values generated by the updated licensees PRA, the differences were investigated through a detailed review of the results from the case runs. This included review of the plant specific assumptions and the minimal cutsets generated by the PRA model for evaluating the impact of the hypothetical inspection findings.

As a result of benchmarking, the SDP notebook now reflects the plant-specific characteristics within the framework of the SDP modeling approach considering the latest plant-specific PRA for the plant.

2. Summary Results from Benchmarking Summary of Benchmarking Results Benchmarking of the SDP Notebook for the Clinton NPP (CLNT) was conducted comparing the order-of-magnitude results obtained using the notebook with that obtained using the plant-specific PRA. Cases for which SDP notebook results were under or overestimates were identified. Three cases of a conservative result by two orders of magnitude (i.e., the significance obtained using the notebook was two colors higher than that to be obtained using the plant PRA) and ten cases of conservative results by one order of magnitude were noted. In addition, two cases of underestimation by one color were noted. A summary of the results of the risk characterization of hypothetical inspection findings is as follows:

5% (2 of 43 cases) underestimation of risk significance 7% (3 of 43 cases) overestimation of risk significance by two orders of magnitude 16% (7 of 43 cases) overestimation by one order of magnitude 72% (31 of 43 cases) consistent risk significance.

Detailed results of Benchmarking are summarized in Table 1. Table 1 consists of eight columns. The first column identifies the components/failed operator actions or the case runs.

The assigned colors from the SDP Rev. 0 worksheets without incorporating any modification from the Benchmarking exercise are shown in the second column. The third column either gives the basic event name in the plant PRA, if applicable, or indicates that the results from a case run were used to obtain the risk achievement worth (RAW) for the out-of-service component or the failed operator action. The fourth and fifth columns respectively show the licensees internal RAW value and the color to be defined based on the RAW values from the latest PRA model. The sixth column presents the colors for the inspection findings based on the Rev. 1 version of the notebook. The Rev. 1 version of the note book was prepared considering the revisions to the Rev. 0 version of the SDP notebook judged applicable following Benchmarking. The seventh column shows the comparison of SDP results vs. the RAW values obtained from the plant PRA. Those cases where SDP results are consistent with the plant PRA are denoted as M for Match. Similarly, O1", O2", and U stand for overestimation by one color, by two colors, and underestimation respectively. The last column provides some clarifying comments identifying the difference in results between the SDP Rev. 1 notebook and the plant PRA, and the applicable rules in obtaining the color of the inspection finding using the SDP notebook.

Table 2 presents a summary of the comparisons between the results obtained using the Clinton Power Station Notebook and the plant PRA. It also showed a comparison of the results using the Rev. 0 and Rev. 1 versions of the notebook. The results show that underestimations by the notebook were significantly reduced through revisions to the notebook implemented as a result of Benchmarking, i.e., it was reduced from 35% to 5%. However, overestimations slightly increased indicating that the notebook will be slightly more conservative. The matches between the notebook and the plant PRA increased from 35% to 72% indicating significant improvement in the Rev. 1 notebook in comparison to the Rev. 0 notebook.

Discussion of Non-conservative Results or Underestimations by the Notebook Two cases of underestimation were noted during the Benchmarking. They were both related to division II of emergency power, i.e., the 4160 VAC and 125 VDC electrical power. The investigation of the minimal cutsets indicated that the major contributor to the risk significance of loss of division II DC bus stems from the scenarios in which the loss of DC bus occurred after transients with failure of HPCS and unavailability of the division I battery charger (due to failure or maintenance). The PRA assumes, albeit conservatively, that the loss of battery charger was the same as the loss of the DC bus since the battery was not sufficient for the 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> mission time; that is, loss of both Division I and Division II DC after a transient with failure of HPCS would result in core damage. Similar scenarios were also dominant contributors to risk significance of the loss of Division II AC bus. As it was determined previously, SDP notebooks would underestimate the scenarios involving the failure of multiple support systems, i.e., losses of DC Divisions I & II, after a reactor trip when the loss of any of the support system was not the cause of trip (a special initiator).

Discussion of Conservative Results by the Notebook Eleven cases of overestimation (four cases by two colors and seven cases by one color) were noted during Benchmarking. We first made some general observations regarding the overestimations by the SDP notebook and then discuss the overestimations by more than one color.

There were some major differences in SDP credit for mitigation capability compared to the unavailability of the equipment in the plant PRA. As an example, RCIC was assigned a credit of 1 ASD train, equivalent to an unavailability of 1.0E-1 whereas in the plant PRA RCIC unavailability was approximately 2.2E-2. The human error probability (HEP) value associated with primary depressurization was 5E-4 where the SDP assigns a credit of 3. Several of these differences have contributed to the cases where the SDP overestimated the PRA results by one color. Such differences are expected per SDP construction and evaluation rules.

An overestimation by two colors was noted for HPCS battery charger, Battery charger for DC division I, and failure of 1 SLC pump during ATWS. These cases are described below:

1. The overestimation of the risk significance of HPCS battery charger and the battery charger for Division I by two orders of magnitude resulted primarily from not crediting recovery actions in the SDP notebook. Failure of a battery charger during normal operation was typically diagnosed prior to resulting in a loss of associated DC bus and actions would be taken to recover from that failure. The SDP notebook will provide a closer match (overestimate by one order of magnitude) to the plant PRA when the actual performance issue is identified and analyzed through the inspection finding process and the recovery action is credited.

2. The overestimation of 1 SLC pump by two orders of magnitude was attributed to the difference in ATWS frequency and the success criteria for the SLC pumps. ATWS was assigned to Row V in the SDP notebook but has frequency of approximately 2E-6/reactor-year in the plant PRA. Also, in the plant PRA, 1 of 2 SLC pumps were needed if the operators can initiate the SLC within a short time. In the SDP notebook, 2 of 2 pumps were assumed needed since the HEP for operator succeeding in a short timewas assumed high.

Changes Incorporated Following Benchmarking Resulting in Updating of Benchmarking Results The results obtained during benchmarking were conservative in most cases resulting in further review of the models in comparison with the detailed cutsets provided by the licensee. Based on review of these cutsets and other related information, modeling in the SDP notebook was substantially revised. This resulted in revised event trees and worksheets for all initiators. The changes made to complete the Rev. 1 notebook are listed in Section 3. Changes following the benchmarking are not separately addressed here.

ATTACHMENT A Table 1: Summary of Benchmarking Results for Clinton CDF = 1.26E-5 at Truncation level 1E-9 RAW thresholds are W = 1.08, Y = 1.79, and R = 8.4 Color Rev. 0 SDP Basic Event Clinton by Mod. SDP Comparison Comments Item Out of Worksheet Names RAW Clinton Worksheet Service Color RAW Color Component HPCS Red HPSYST1SYM 11.2 R1 R2 O1 HPSYST1SYM and case run (Also a case run) 1 CRD Pump Green NRDBPRETRM 1.04 G G M RHR Train (SPC Green R2B001BHXP 1.29 W W M mode) EDG-3 Yellow APDG1KCCBC 2.38 Y Y M RCIC White H22N054LSH 5.5 Y Y M PCS White Case Run 3.9 Y Y M The dominating (Steam) CDF = factor here is the 2.88E-5 loss of room cooling due to CCFs.

PCS White Case Run 1.62 W Y O1 (TPCS initiator) CDF =

7.8E-6 1 SRV fto White Case Run 1.0 G W O1 3 SRVs fto White YMSSRVMRVO 3.41 Y Y M 1 SRV ftc Green YMSSRVXRVC 3.64 Y Y M

ATTACHMENT A Color Rev. 0 SDP Basic Event Clinton by Mod. SDP Comparison Comments Item Out of Worksheet Names RAW Clinton Worksheet Service Color RAW Color LPCS Green BRABCCCMPS 1.02 G W O1 RHR - pump A or Green RH2RHRBSYH 1.12 W W M B

RHR - pump C Green R3PREVTSYM 1.01 G G M One SLC pump Green Y1SC20CXXH 1.01 G Y O2 Both SLC pumps Yellow SJOINTXSYM 3.0 Y Y M EDG 1: Div. I and White APDG1KACBC 1.86 Y Y M Failed the Div. I LOOP Kills CV equipment and CV in LOOP worksheet.

EDG 2 White 1.79 W W M 4160 AC Div. I White Case Run 19.7 R R M It also considers RCIC for Room the failure of RCIC Cooling due to loss of room cooling.

4160 AC Div. II Yellow Case Run 14.6 R Y U 4160 AC Div. III Not Modeled Case Run 11.2 R R M Same as HPCS.

Safety-related Yellow 20.7 R R M DC Div. I Safety-related White 10.6 R Y U Also impacts the DC Div. II ISLOCA injection path.

ATTACHMENT A Color Rev. 0 SDP Basic Event Clinton by Mod. SDP Comparison Comments Item Out of Worksheet Names RAW Clinton Worksheet Service Color RAW Color Safety-related Red 11.2 R R M The impactwas DC Div. III or IV loss of HPCS.

Battery of safety- Yellow D1DC01EBYM 3.11 Y Y M related DC Div. I Battery for DC Not Modeled D1DC02EBYM 1.5 W W M Div. II Battery for Div. III Not Modeled DCS001DBYM 2.33 Y Y M BOP Battery Not Modeled 1.3 W W M HPCS Battery Not Modeled D1S001EBCD 1.67 W R O2 Charger CONSERVATIVE Battery charger Yellow D1S01EBCD 1.67 W R O2 of safety-related DC Div. I Battery Charger Not Modeled D1DC26EBCD 1.58 W W M for non-safety DC D1DC25EBCD Loss of White IEYLOSSIA 1.62 W W M Instrument Air Loss of an air Green IEYLOSSIAI: 1.0 G W O1 Compressor Case run as an initiator Plant SW pump Green IEYLOSSWI: 1.0 G G M (WS) Case run as an initiator

ATTACHMENT A Color Rev. 0 SDP Basic Event Clinton by Mod. SDP Comparison Comments Item Out of Worksheet Names RAW Clinton Worksheet Service Color RAW Color Train A of Yellow XSX063AMVO 1.18 W Y O1 shutdown SW (SX)

Train B of White XSXSYS2SYH: 1.99 Y Y M ISLOCA shutdown SW Contribution Not contribution not (SX) Included included Train C of Red XSXSYS3SYH 2.67 Y Y M shutdown SW (SX)

ONE TBCCW White WWT1PBPMPM 1.03 G G M Pump Operator Actions DEP Red GADSMANSYH 117.73 R2 R2 M LVL Control Yellow YATWSLCHH 2.56 Y Y M INH Yellow 1.0 G G M Failure to inhibit in ATWS will not results directly in core damage.

REC4 Green YOS0TD4SWH 1.21 W W M Cont. Vent Not Modeled Case Run 1.57 W W M Trip Recirc. Yellow YRECIRCRDY 2.18 Y Y M Pump

ATTACHMENT A Color Rev. 0 SDP Basic Event Clinton by Mod. SDP Comparison Comments Item Out of Worksheet Names RAW Clinton Worksheet Service Color RAW Color Align RHR for Yellow Case Run 2.64 Y R O1 SPC mode 3.3E-5/

1.26E-5 Notes:

1. Clintons RAW for internal events, average maintenance case.

2. 4160 AC Div. II is needed to open MOVs to use fire pumps to inject into RPV through WS to SX.

ATTACHMENT A Table 2: Comparative Summary of the Benchmarking Results SDP Worksheet SDP Worksheet Modified Number of Cases Percentage Number of Cases Percentage SDP: Non-Conservative 15 35 2 5 SDP: 1 order 5 11.5 7 16 Conservative by 2 orders 1 2.5 3 7 SDP: Matched 15 35 31 72 SDP: Not Modeled 7 16 0 0 Total 43 100 43 100 3 ADDITIONAL PROPOSED MODIFICATIONS TO SDP WORKSHEETS 3.1 Specific Changes to the Rev. 0 SDP Worksheets for the Clinton SDP A number of changes were made to the Clinton worksheets. Changes made during and after the plant onsite visit are summarized below and are contained in the updated notebook.

Table 1:

1. Moved IORV to Row I and changed it to SORV with a footnote indicating that the total frequency in the PRA is 0.1 per reactor-year.

2. Added loss of IA in Row I with a footnote reflecting the initiating event frequency in the PRA is 9.1E-2 per reactor-year.

Table 2:

1. Included several minor comments made by the licensee on the system dependencies.

2. Added a footnote saying that Div IV DC power would provide redundant actuation signal for HPCS.

3. Added ECCS room cooling with major components as chillers and fans with support provided by Divs I, II, and III of SX and 480 VAC.

4. Added a footnote indicating that each EDG is equipped with a day tank with 700 gallons of fuel oil sufficient for 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> of EDG operation and each EDG has a separate fuel transfer train.

5. Added a footnote reflecting that battery chargers can carry SI loads without a battery.

However, the batteries by themselves (without battery chargers) are not credited in the PRA since they would not last the 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> mission time. Also note that the battery duration during SBO is 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />.

Table 3.1:

1. Added CV and redid the event trees.

2. Changed the DEP success criteria to 3/16 SRVs.

3. Added a footnote specifying that there are three paths for containment venting. The HEP value associated with these paths are 0.1, 0.1, and 0.5 per demand. Due to human action dependency resulting from diagnostics, a total credit of 2 is assigned.

Table 3.2:

1. Incorporated similar comments to table 3.1.

2. Added a new heading to the event tree called CCFRM reflecting loss of all ECCS room cooling due to common cause failure. Added a footnote saying the probability of this event is 7.0E-5 and could be recovered. Assigned a CCFRM credit of 5.

Table 3.3 :

1. Modified the event tree requiring CV and 1/2 CRD pumps when the RCIC is the only means of injection.

Table 3.4:

1. Modified event tree to require CV when CHR has failed and required 2/15 remaining SRVs for depressurization. Removed the credit for PCS and changed IORV to SORV.

Table 3.5:

1. Changed DEP to 3/16 SRVs and maintained HEP credit of 2.

Table 3.6:

1. Deleted the mitigation capability for early containment failure (EC) and added a footnote that it would be unlikely in Clinton.

Table 3.7:

1. Removed the credit for CRD, required CV, and modified the event tree accordingly.

2. Dropped the credit for REC4 from 2 to 1 and added a footnote explaining that this credit includes the success of both DC load shedding and recovery of AC power.

Table 3.8:

1. Change the credit for SLC from 1/2 to 2/2 SLC pumps and reduced the credit to 1 train.

Table 3.9:

1. Modified the LDC event tree similar to TPCS event tree.

2. Included in footnote 1 that the condensate system is credited for steam condensing mode as a part of RHRSPC.

Table 3.10:

1. Modified the LSW event tree similar to TPCS event tree.

Table 3.11:

1. Added a new event tree and a worksheet for LRAT (it would be similar to LSW). The loss of RAT causes the divisional loads to be transferred to ERAT. If the transfer is successful, then manual shutdown of the plant will be pursued which will result in loss of SW.

2. Included in the first footnote that the LRAT frequency is 4.0E-2 per reactor-year.

Table 3.12:

1. Modified the LIA event tree similar to TPCS and added CV to the event tree.

3.2 Generic Changes in IMC 0609 for Guidance to NRC Inspectors None.

3.3 Generic Change to the SDP Notebooks None.

4 DISCUSSION ON EXTERNAL EVENTS As analyzed by the licensees updated PRA models, the core damage frequency estimates for internal initiators was 1.2E-5 events per reactor-year. Clinton does not have an integrated external event PRA. The licensee noted that the flood contribution was negligible and was estimated in the range of 1.E-8 to 1.E-9 per reactor-year.

5. LIST OF PARTICIPANTS Adel El-Bassioni US NRR/NRC Mike Parker US NRC/Region III Mark Reinhart US NRR/NRC Peter Wilson US NRR/NRC M. Ali Azarm BNL Robert Buell INEEL Thomas Daniels Exelon/Erin Engineering Terry Lanc Exelon Xavier Polanski Exelon John Steinmetz Exelon