ML021220239
| ML021220239 | |
| Person / Time | |
|---|---|
| Site: | Columbia  |
| Issue date: | 04/28/2002 |
| From: | Reinhart F NRC/NRR/DSSA/SPSB |
| To: | Carpenter C, O'Reilly P NRC/NRR/DIPM, NRC/RES/DRAA/OERAB |
| WILSON J, NRR/DRIP/RGEB 415-1108 | |
| References | |
| Download: ML021220239 (17) | |
Text
April 28, 2002 NOTE TO: Cynthia Carpenter, 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, Chief/Signed by M. Caruso for Licensing Section Probabilistic Safety Assessment Branch Division of Systems Safety and Analysis Office of Nuclear Reactor Regulation
SUBJECT:
RESULTS OF THE COLUMBIA GENERATING STATION SDP PHASE 2 NOTEBOOK BENCHMARKING VISIT During December, 2001, NRC staff and a contractor visited the Columbia Generating Station (CGS) to compare the CGS Significance Determination Process (SDP) Phase 2 notebook and licensees risk model results to ensure that the SDP notebook was generally conservative.
CGSs PSA did not include external initiating events 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 CGS were also compared with the licensees risk model. The results of the SPAR model benchmarking effort will be documented in a separate trip report to be prepared by the Office of Research.
In the review of the CGS SDP notebook, it was found that some changes to the SDP worksheets were needed to reflect how the plant is currently designed and operated. Forty hypothetical inspection findings were processed through the SDP notebook. Results from this effort indicated that the total risk impacts modeled in the SDP notebook were underestimated by 18 percent, overestimated by 40 percent, and adequately estimated by 42 percent. The reviewers found that if thirteen fixes were made to the SDP notebook, the results would be 10 percent underestimation and 42 percent overestimation of risk impacts.
Attachment A describes the process and results of the comparison of the CGS SDP Phase 2 Notebook and the licensees PSA.
If you have any questions regarding this effort, please contact Peter Wilson.
Attachments: As stated CONTACT: P. Wilson, SPSB/DSSA/NRR 415-1114
April 28, 2002 NOTE TO: Cynthia Carpenter, 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, Chief/Signed by M. Caruso for Licensing Section Probabilistic Safety Assessment Branch Division of Systems Safety and Analysis Office of Nuclear Reactor Regulation
SUBJECT:
RESULTS OF THE COLUMBIA GENERATING STATION SDP PHASE 2 NOTEBOOK BENCHMARKING VISIT During December, 2001, NRC staff and a contractor visited the Columbia Generating Station (CGS) to compare the CGS Significance Determination Process (SDP) Phase 2 notebook and licensees risk model results to ensure that the SDP notebook was generally conservative.
CGSs PSA did not include external initiating events 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 CGS were also compared with the licensees risk model. The results of the SPAR model benchmarking effort will be documented in a separate trip report to be prepared by the Office of Research.
In the review of the CGS SDP notebook, it was found that some changes to the SDP worksheets were needed to reflect how the plant is currently designed and operated. Forty hypothetical inspection findings were processed through the SDP notebook. Results from this effort indicated that the total risk impacts modeled in the SDP notebook were underestimated by 18 percent, overestimated by 40 percent, and adequately estimated by 42 percent. The reviewers found that if thirteen fixes were made to the SDP notebook, the results would be 10 percent underestimation and 42 percent overestimation of risk impacts.
Attachment A describes the process and results of the comparison of the CGS SDP Phase 2 Notebook and the licensees PSA.
If you have any questions regarding this effort, please contact Peter Wilson.
Attachments: As stated CONTACT: P. Wilson, SPSB/DSSA/NRR 415-1114 DISTRIBUTION: SPSB r/f PWilson MReinhart G:\\SPSB\Wilson\CGSbench.wpd Accession#ML021220239 OFFICE SPSB SC:SPSB NAME PWilson:nyc Mreinhart/Signed by M. Caruso for DATE 04/23/02 04/28/02 OFFICIAL RECORD COPY
SUMMARY
REPORT ON BENCHMARKING TRIP to the Columbia Nuclear Power Plant Unit 1 (December 11-13, 2001)
J. C. Higgins Energy Sciences and Technology Department Brookhaven National Laboratory Upton, NY 11973-5000 December 2001 Attachment A
Table of Contents Page
- 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
- 2. Preparation for Benchmarking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.1 Main Changes to Notebook Prior to Onsite Visit . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.2 Questions for Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.3 Proposed Changes to Worksheets in Maintenance Phase . . . . . . . . . . . . . . . . . . . 2
- 3. Summary Results from Benchmarking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
- 4. Additional Proposed Modifications to SDP Worksheets . . . . . . . . . . . . . . . . . . . . . . . . . 10 4.1 Specific Changes to the Rev 0 SDP Worksheet for CGS . . . . . . . . . . . . . . . . . . . 10 4.2 Generic Change in 0609 for Inspectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 4.3 Generic Change to the SDP Worksheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
- 5. Discussion on External Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
- 6. Persons Participating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
- 7. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 List of Tables Page Table 1. Comparison of Component Sensitivity Calculations between Phase 2 Worksheets and Columbia RAWs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Table 2. Comparative Summary of the Benchmarking Results . . . . . . . . . . . . . . . . . . . . . . . 9
-ii-
- 1. Introduction This report compares the NRC Risk Informed Inspection Notebook, developed by BNL, and the licensee risk model for the CGS NPP to ensure that the Notebook is generally conservative. The benchmarking was performed after the worksheets were revised to include the appropriate licensee comments and recommendations.
- 2. Preparation for Benchmarking The Columbia inspection notebook was originally prepared in August, 2001. The Columbia notebook was reviewed prior to this benchmarking visit in order to identify potential changes that may be needed. A few changes were made (Section 2.1) and some areas were identified for clarification onsite (Section 2.2). Some other areas were identified for changes that will be made in Revision 1 of the notebook.
2.1 Main Changes to Notebook Prior to Onsite Visit C Dropped credit for LDEP from LOSW worksheet since it is lost on LOSW and added clarifying note.
C Changed CV line on Table 2, last column, to indicate not to consider LOSW worksheet.
2.2 Questions for Site
- a. What is the total PRA IE frequency for those transients that involve a loss of PCS?
- b. Provide a discussion of the current containment venting arrangement, paths, valves, and support systems. What is the PRA HEP for operators fail to vent?
- c. What is affected by a finding in the CSTS: PCS, RCIC, HPCS, LPCI? May need to adjust Table 2 for CSTS. Now has all.
- d. CAS affects PCS, CIA, CSTS, & CV. How significantly does a finding in CAS affect these systems? Revisit Table for CAS re the CSTS and CV dependency. Note that just CV would give a Yellow.
- e. CIA affects PCS and SRVs. The SRV dependency would give W bcr. Probably OK but may be too conservative. Check with licensee.
- f. Does loss of a single SW pump (A or B) cause a loss of LPCI, RHR, LPCS EDG, RBEC or is there a cross connect?
- g. What support does RCC provide to PCS? What is the effect of loss of RCC? Clarify this in Table 2. Table now shows no system with RCC as support.
- h. Need to clarify the impact of loss of RBEC and update note 16 of Table 2.
The licensee provided answers to all the above questions during the site visit.
2.3 Proposed Changes to Worksheets C Add notes to Table 2 for RHR pumps A, B, & C, and for RCC.
- 3. Summary Results from Benchmarking The onsite visit was conducted by a BNL representative, an NRC headquarters PRA representative and the Region IV SRA. We met with the Columbia PRA staff for two and one half days onsite to perform the benchmarking activities.
The initial activities included reviewing with the licensee the updates made to the CGS notebook as described in Section 2.1 above, and provided an updated copy of the Notebook to CGS. We then provided CGS with the questions in Section 2.2. CGS researched the related information and provided answers to the team. CGS also gave the team a few added comments to the Notebook, which were valuable and improved the notebook. These were incorporated before beginning the benchmarking activities. Notable changes made, based on the licensees comments and justifications, were:
Adjustment of the initiating event (IE) frequencies to those currently used in the CGS PSA and revision of Table 1 to agree with the new IE frequencies.
- 1. A few minor changes to Table 2.
- 2. Added credit to all worksheets for LDEP after successful HPCS and failure of CHR and CV.
- 3. On all worksheets, removed the necessity for LDEP after successful RCIC, failure of CHR and successful CV on several worksheets.
- 6. Changed the credit for DC load shed from 3 to 2.
- 9. Added credit for PCS on LODC1 and LODC2 worksheets.
- 10. Added a worksheet for loss of the control and service air (CAS) system.
- 11. Obtained the current detailed system description information on the CV, CAS, and CJW systems and incorporated this into the notebook.
The licensee provided updated information to the team based on the CGS PSA, Rev. 4.2, dated 6/22/01. This was supplemented by additional PSA information provided in response to team questions throughout the visit. Information included: definition of basic events, cutsets, RAW values, system design information, analysis assumptions and results, and event trees.
The team computed the break points in RAW values for the different SDP colors based upon a current PSA total internal events CDF of 2.25 E-5 events/reactor-year. The team had pre-selected a fairly large list of components and human actions, as listed in Table 1 below, that would be evaluated for the effect of having the component or human action fail. Prior to the site visit, the team developed the color corresponding to failure of each item. This list of items was modified slightly onsite. We then used the latest revised version of the notebook to develop the color corresponding to failure of each item and compared that to the color that would be implied by the items RAW value from the PSA.
In developing the colors from the notebooks, the team evaluated all sequences in each worksheet that contained the item (component or human action). A number was obtained for each re-evaluated sequence. We then used a counting rule to cascade lower value sequences to higher value ones as follows. For example, three sequences of value 8 (shorthand for an estimated sequence frequency of 1 E-8 events/reactor-year) were equivalent to one sequence of value 7.
Likewise 3 sequences of value 7 (3-7s) were equivalent 1 sequence of value 6(1-6). Also, 3-6s were equal to 1-5, and so on. Colors were developed as follows:
Sequences of value 7, 8, and higher Green Sequences of value 6 White Sequences of value 5 Yellow Sequences of value 4 or less Red When the above described counting rule was needed to obtain a color rather than a direct correlation from a sequence, then in Table 1 we note that it was obtained by the counting rule or bcr.
Table 2 provides a summary of the benchmarking results. The teams initial quantification of the Notebook had seven non-conservative items and 16 that were more conservative than the colors that were based on the RAW values from the licensees PSA. The teams final quantification of the Notebook had four non-conservative items and 19 that were more conservative than the colors that were based on the RAW values from the licensees PSA.
For two of the non-conservative findings (LPCS and 1 SRV fail to close) the team found that their importance in the licensees model comes from internal flooding sequences. When the contribution from internal flooding is removed, then we obtain a match between the notebook and the plant PSA. More specifically, we examined the dominant cutsets containing LPCS components and SRV (FTC) and found that they were all flooding initiators that flooded ECCS equipment rooms, areas R404 & R 405. This resulted in failure of: condensate, FW, recirculation pumps, CAS, RHR A, and HPCS. It also caused turbine trip/reactor trip. When this is combined with a stuck open SRV, one can see the increase in the importance of LPCS and SRV (FTC). We have noted this in Table 2 of the Notebook.
BNL notes that this was the second BWR-5/6 plant that has had a benchmarking visit and several useful Columbia plant systems and PSA insights were obtained from the licensee. After the first BWR-6 visit to Perry, it was recommended that, a benchmarking trip should be performed at another BWR-5/6 in order to consolidate the insights and enable the production of improved SDP notebooks for all BWR-5 and 6 plants.
Table 1: Comparison of Component Sensitivity Calculations between Phase 2 Worksheets and Columbia RAWs (CDF = 2.25 E-5; RAW splits - 1.04, 1.44, 5.44)
Truncation level 1 E-10 SDP Color Item Out of Work- Colum. by Mod. SDP Service sheet Columbia RAW1 Colum Worksheet Component Color Basic Event ratio RAW1 Color Comments HPCS Rbcr HPSP-MD- 5.2 Y Rbcr conserv.
1R2LL EDG-3 Y EACENG- 3.3 Y Y EDG3-W2D3 RCIC W RCIP-TD-1R2LL 3.13 Y Y 1 SRV fto W bcr 1 G W conserv.
2 SRVs fto Y MC 1.18 W Y conserv (OVERPR) 1 SRV ftc G 2.15 Y G non-conserv note 1 LPCS G LPSP-MD- 1.6 Y G non-1R2LL conserv.
note 1 RHR- pump R RHRPMD- 11 R Rbcr A 2AR2LL RHR-pump R RHRPMD- 14 R Rbcr B 2BR2LL RHR-pump G RHRPMD- 1 G G C 2CR2LL RHR HX R 2.94 Y R conserv 1 CV valve Y 1.03 G W conserv One SLC Y SLC-PMD 1.18 W Y conserv pump 1AR2XX SDP Color Item Out of Work- Colum. by Mod. SDP Service sheet Columbia RAW1 Colum Worksheet Component Color Basic Event ratio RAW1 Color Comments EDG 1 or 2 Y EACENG- 4 Y Y note 6 EDG2-W2D2 (or 1) 4160 AC R ~760 R R Div. 1 4160 AC R ~4000 R R Div. 2 CIA item W bcr SRV 1 G W conserv.
CAS item --- 1 CAS 1.1 W Y conserv.
compressor TSW item Y 1 TSW train 1.4 W Y conserv.
One DW-SP W note 8 Y vacuum breaker drywell floor R D 111 R R note 8 seal DC-Div 1 R 42 R R DC-Div 2 R 40 R R DC Battery R EDCB1-EB1- 4.53 Y R conserv.
A 1W2D1 DC Battery R EDCB1-EB1- 2.83 Y R conserv.
B 2W2D1 DC charger R EDCC1-EC1- -- -- R A 1W1D1 DC charger R EDCC1-EC1- 4.92 Y R conserv.
B 2W1D1 SW A R 19 R R SW B R 36 R R PCS Y/G 11.6 R Y non-conserv.
see note SDP Color Item Out of Work- Colum. by Mod. SDP Service sheet Columbia RAW1 Colum Worksheet Component Color Basic Event ratio RAW1 Color Comments RPT-1 train G -- -- G RPT-both Y 1.13 W Y conserv.
trains all of CJW -- 13.9 R Y non-conserv. -
note 9 Operator Actions DEP R ADSHUMNSTA 97.5 R R RTH3LT (non ATWS)
SW cross-tie W RHRHUMNSW 2.16 Y Y CRTIELL SLC Y 1.14 W Y conserv.
CV R VENTFAIL 1.21 W Y conserv.
RLOOP3H G NRAC3 1 G G RLOOP6H G NRAC6 1.72 Y Y RLOOP24H Y NRAC24 2.27 Y R conserv.
DC load G RCIHUMN- 1.2 W W shed LOADSHED INH/LC Y AI 1.07 W Y conserv.
RHR SP R RHRHUMN 64.4 R R cooling SPCOOLLL Notes:
- 1. LPCS and one SRV (fail to close) were non-conservative. The cutsets were examined and the full non-conservatism was due to internal flooding sequences that are not addressed in the current notebooks.
- 2. Columbia RAW for internal events, average maintenance case.
- 3. The Delta CDF used in RAW value calculations represented the change in CDF due to the component being out of service for 1 year.
- 4. The subscript bcr means by counting rule.
- 5. For a component such as a pump, we examined the RAW values for the basic events both for failure to start and failure to run, and either selected the highest (more conservative) value here, or used a synthesized RAW value separated calculated by the licensee that included all failure modes.
- 6. For those items where the basic event column is blank either we were unable to identify a PSA modeled basic event that was equivalent or the licensee used a synthesized RAW value separated calculated that included all failure modes.
- 7. Regarding the evaluation of EDG Div. 1 and 2, we used newly developed SBO and LOOP/EDG worksheets. Then for the SBO worksheet we increased the SBO initiating event frequency by two orders of magnitude for the evaluation and failing any components affected by loss of the EDG.
- 8. For PCS, if we assume a total loss, then the color is Y, but if we just assume a degradation and reduce credit from 3 to 2, then we get a G. Further the licensee provided the team with a best estimate RAW value for PCS, but noted that they do not model PCS per se. Rather, as with most other sites, they have a much more detailed model. Also they take considerable credit for recovery of PCS components, which increases the overall importance of PCS.
- 9. Both the CGS PSA and the notebook model early containment control (EC) success as Passive operation of SP: 1 of 2 vacuum breakers in 9 of 9 lines remain closed and drywell floor seal intact. The PSA uses a fault tree to model the 18 vacuum breakers and combines this with seal reliability estimates to obtain one event (D) with one failure probability number used in the PSA. Thus with their PSA model they cannot easily obtain a RAW for one vacuum breaker, but they can obtain a RAW for the D function. Also the Yellow for one vacuum breaker in the notebook is somewhat conservative, since it assumes that the next failure is in the same line as the one with the inspection finding.
- 10. Loss of CJW (jacket cooling water system) will fail 3 of 4 air compressors of the CAS system.
- 11. We were not able to obtain RAW values for three items (one DW-SP vacuum breaker, DC charger A, and one train of RPT). Thus, these items are in Table 1 but are not included in the summary of Table 2.
Table 2: Comparative Summary of the Benchmarking Results SDP Worksheet SDP Worksheet Modified Number of Cases Percentage Number of Cases Percentage SDP: Less 7 18 4 10 Conservative SDP: More 15 40 17 42 Conservative SDP: Matched 16 42 19 48 Total 38 100 40 100
- 4. Additional Proposed Modifications to SDP Worksheets 4.1 Specific Changes to the Rev 0 SDP Worksheet for CGS A number of changes were made to the CGS worksheet. Ones made before the onsite visit are noted is Section 2 above. A number of additional changes, made during and after the plant onsite visit, are summarized in Section 3 above and are contained in the updated notebook. This update has been supplied in draft form to the NRC Region III Senior Reactor Analyst.
4.2 Generic Change in 0609 for Inspectors Some comments on the 0609 usage rules were provided to the Region IV SRA for forwarding on to NRR as part of the 0609 comment process.
4.3 Generic Change to the SDP Worksheet None.
- 5. Discussion on External Events As analyzed by the licensees updated PSA models, the core damage frequency estimates for internal initiators was 2.25 E-5 events/reactor-year. Columbia does not have an integrated external event PSA. The licensee noted that their IPEEE study for fire, floods, and seismic events does not have sufficient information to provide insights to potential changes in color evaluation based on consideration of external events.
- 6. Persons Participating NRC:
Pete Wilson Troy Pruett BNL:
Jim Higgins INEEL:
Bob Buell Columbia NPP:
Albert Chiang Lichung Pong
- 7. References Columbia PSA, Rev. 4.2, dated 6/22/01.