NUREG-2230, Comment Resolution - NUREG-2230 Final
| ML19291A339 | |
| Person / Time | |
|---|---|
| Issue date: | 07/31/2019 |
| From: | Stroup D NRC/RES/DRA/HFRB |
| To: | |
| D. Stroup, NRC/RES/HFRB, 301-415-1649 | |
| References | |
| EPRI 3002016051, NUREG-2230 | |
| Download: ML19291A339 (46) | |
Text
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date RC-1 Page 3-14, Sentence 1 Add a parenthetical at the end of the sentence to provide Revised as suggested. 8/1/19 more direction of the applicability of the approach. "...
suppression rate (Interruptible, Growth, transient, etc.)."
RC-2 Figures 4-3 and 4-4 Modeling doesnt match the summary description Key Agreed. There was an inconsistency between how 8/5/19 Finding on pdf page 9. the data was averaged versus how the test data was developed to be applied during fire modeling.
Revised the key findings to clarify the parameters for fire modeling. These bullets now read:
Based on an analysis of relevant experimental evidence, the HRR timing profile for interruptible fires can be modeled in one of two ways (see Section 4.2.2):
o Using a pre-growth period of 9 with a negligible HRR, 7 minute time to peak, 5 minutes at steady state, and a 13-minute decay period.
o The addition of a pre-growth period to the timing profile prescribed in NUREG/CR-6850 A pre-growth period of 4 minutes with a negligible HRR 12 minute time to peak, 8 minutes at steady state, and a 19-minute decay period.
Page 1 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date For growing fires, the HRR timing profile in NUREG/CR-6850 is recommended (12 minute time to peak, 8 minutes at steady state, and 19-minute decay period). See Section 4.2.4.
RC-3 Figure 5-11 The word fire is over top of the y axis labels. Shrank text to fit on one line 7/31/19 RC-4 Page 6-5 Why is 4 minutes used versus 8 minutes for the pre-growth See updated discussion in Section 4.2 (and 8/5/19 period? comment RC-2)
RC-5 Section 4.1.1 In Section 4.1.1, an interruptible fire is said to have a growth Revised Section 4 to better separate the process 8/5/19 period up to 8 minutes. Later in in Section 4.2, 4 minutes is of analyzing the fire test data with the fire modeling
& recommended. As no criteria is provided to determine what recommendations. The analysis of data is in Section 4.2 value to use, this introduces unclear and ambiguous Section 4.1. The interpretation of the results and guidance. The pre-growth period should be defined and the fire modeling values are in Section 4.2. The used like the growth, steady-state, and decay periods. A revision for Section 4.2 now describes that:
single recommended value should appear in all sections a.) When modeling using the experimentally and examples within the NUREG. HRR profile using times presented in Section 4.1.1 a pre-growth period duration of 9 minutes may be used.
b.) When maintaining modeling results using the NUREG/CR-6850 HRR profile durations, a pre-growth period with a duration of 4 minutes may be used to ensure that the time to peak does not become greater than that of the experimentally derived time.
See the revised discussion in Section 4.2.1.
Page 2 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date WebEx 5.3.3.2 Does MCR indication only apply to an annunciated alarm? Yes. Operator response to audible alarms is based 08/22/2019 (SL-1) Would it apply to other types of indications in the MCR? on both the attention-getting and importance of the audible cue and training on the associated Alarm Response Procedures, which requires the operators to investigate the source and confirm the validity of the alarm.
While it is recognized that operators do continually scan the board for changes and review the status of equipment upon a shift change, it is not believed that the speed and reaction to a warning light would be the same as an annunciated alarm.
NEI-1 Statement from report: Key Findings: The typical HRR profile for The Key Findings have been updated to list all 8/5/19 interruptible fires is a pregrowth period with a negligible HRR for applicable modeling profiles. See Comment RC-2 up to 8 minutes, 12minute time to peak, 8 minutes at steady state, and a 19minute decay period (see Section 4.1.1).
viii Comment: The growth profile for interruptible fires discussed in Section 4.1.1 is pregrowth for 8 minutes, 7 minutes time to peak, 5 minutes steady state, and 13 minutes decay. If both curves are acceptable for use, then both should be listed as key findings.
NEI-2 Equipment trouble alarms in the main control room (MCR) due to See response to SL-1 8/22/19 fire will occur in the early stages of the fire development.
32 Suggest clarification if other notification to the MCR is acceptable other than trouble alarms.
NEI-3 Table 32 and Table 33 Agreed, changed the title of Table 3-2 to read 8/12/19 39 & 310 It may be worth reiterating that the data used for Table 32 is the Interruptible and growing split fractions (2000-20002014 data because the next page shows Table 33 which 2014) and Table 3-3 to Electrical ignition source Page 3 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date uses the 19902014 data for the nonsuppression curve, and the probability distribution rate of fires suppressed per number of interruptible vs growing fires is not equivalent unit time (1990-2014) between the tables.
NEI-4 3.5.2 NSP Floor for the MCR The NSP floor discussion will remain in NUREG- 8/12/19 2230 since this report will be published first. In NUREG-2178 Vol 2 the section on the NSP floor 312 This data is presented in both NUREG2230 and NUREG 2178 Vol. will be eliminated and the reader will be pointed
- 2. Suggest removal from one of the NUREGs replacing it with a back to NUREG-2230 for the technical basis.
simple reference to the other.
NEI-5 Revised as: 08/13/19 MCR will use the interruptible and growing suppression rates For example, a cabinet fire in the MCR will use the presented in Section 3.5 for fire durations in excess of 18 minutes. Control Room suppression rate up to a value of 314 & 315 Additional clarification could be provided to explain that this 1E-03, after which the Interruptible and Growing crossover occurs at the time on the interruptible or growth curve suppression rates presented in Section 3.5.1 will where NSP=1.00E3 be used for fire durations in excess of 18 minutes.
NEI-6 For example, a fire in a cabinet located within the MCR will use Redundant with comment RC-1 8/12/19 the interruptible and growing suppression rates presented in Section 3.5 for fire durations in excess of 18 minutes.
314 Add a parenthetical at the end of the sentence to provide more direction of the applicability of the approach. "... bin specific suppression rate (Interruptible, Growth, transient, etc.)."
NEI-7 Redundant with comment RC-2 7/31/19 Chapter 4 Figures 43 and 44 Page 4 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date The modeling does not match the summary description key finding on page 12.
NEI-8 A period of up to 8 minutes with no measurable HRR may be See resolution to RC-5 8/5/19 included prior to the period of fire growth. If included, this pre growth phase must be reflected in any calculations of the time to damage, time to detection, and time to suppression.
Here, an interruptible fire is said to have a growth period up to 8 Section 4.1.1 minutes. Later in in Section 4.2, 4 minutes is recommended. As no criteria are provided to determine what value to use, this introduces unclear and ambiguous guidance. The pregrowth period should be defined and used like the growth, steadystate, and decay periods. A single recommended value should appear in all sections and examples within the NUREG.
NEI-9 The growth profile described in NUREG/CR6850 is recommended Agree, we have revised Section 4 to explain the 8/5/19 for modeling growing fires. timing profiles for both growing and interruptible If the use of the NUREG/CR6850 curve is recommended, then it fires. The key findings and summary sections have 47 also been updated.
may be clearer to identify this in section 4.1.2. with the full explanation as is in section 4.2. As it reads right now it appears the NUREG is suggesting this as a new curve, only to say it is not recommended later.
NEI-10 Throughout Reviewed EPRI 3002005302. This text is from 7/31/19 Section 2.5 to discuss the difference between a The fire events database (EPRI 3002005302) that is potentially challenging (did not damage external Chapter 5 referred to in NUREG-2230 indicates that: A distinction is targets, but under alternate could have) and a now made between fires that were capable of damaging a challenging fire (a fire that did have observable PRA important component and those that were not capable effects). Fire events are screened if they are Page 5 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date but, in the absence of automatic or manual suppression, outside of the FPRA scope (administration might have become capable. building, warehouse, cafeteria, etc.). The remaining set of fire events is screened against the However, this distinction appears to simply remove fires that rules in EPRI 1025284 and carried forward in EPRI are outside of the protected area. This screening criteria do 3002005302. The fire events are included if they not properly characterize whether or not a fire has any meet the potentially challenging (or greater) possibility of becoming an initiator that affects FPRA criteria. It is true that screening criteria is components. For example: irrespective if the component(s) have the potential
- 1. The fire could affect components that are in the to result in an initiator. This information is difficult to PRA but never be expected to become an initiator, discern from the event narrative. Likewise, we in which case it would simply become another type agree there are components that may result in an of unavailability. initiator, but may not impact PRA components.
This issue cannot be consistently addressed given On the other hand, the fire could be an initiator, but would the current information collected on fire events.
never be expected to affect FPRA components, and there Please see EPRI 3002016053 for a discussion on are examples of this in the FPRA database. when a conditional plant trip probability can be used for scenarios that are unlikely to result in a plant trip.
NEI-11 Very low: 0.5 - Compartment is subject to controls and The values have been updated to be consistent 08/15/2019 procedures that result in a factor less than a low rating level with the final version of the FAQ.
519 The stated value of 0.5 for very low is incorrect. Per the FAQ 12 0064 closure memo, this value is 0.3.
Page 6 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date NEI-12 See comment above (NEI-11) for the correction on 8/12/19 the very low factor.
Per the text on Page 5-19:
These rating levels are used to make an estimation These modifications are the only exceptions that may be made of the probability that personnel would be present and are done so for the purposes of estimating personnel in a compartment and therefore capable of detection. detecting a fire. Note, the changes described 519 Clarification should be provided whether this only impacts below apply only to estimating the numerical NUREG2230 event tree calculations or also alters existing FAQ probability that personnel are expected to be in a 120064 transient analysis using 0.3 and 3 for very low and compartment. The influencing factors assigned for medium. the purposes of apportioning the transient ignition frequency should be maintained for estimating the probability of personnel detection. No changes may be made to the use of the influencing factor values as applied to the transient ranking scheme as described in NUREG/CR-6850 and FAQ 12-0064.
NEI-13 For occupancy, the rating level associated with medium was See NEI-11 8/12/19 revised from 3 to 5.
521 This discussion should also include the change from 0.3 to 0.5 for very low. This applies to the maintenance factor as well.
NEI-14 Line 4: Recalling Figure 22, almost 50% of the fires that have Agree. Replaced almost 50 with 55% 8/1/19 522 occurred in electrical cabinets have been detected by plant personnel Page 7 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date Figure 22 shows that 55% of the fires that have occurred in electrical cabinets have been detected by plant personnel.
NEI-15 Figure 6-1 reviews the NUREG/CR-6850, 8/12/19 Appendix P approach and does not include the smoke detection ineffectiveness.
As presented in Figure 6-2, smoke detection unreliability and ineffectiveness are included in the detection failure calculation under an OR gate, resulting in an addition of the two terms.
Section 6.1: First Interruptible: 8.41E02 The value for the First Interruptible detection step Treatment of automatic smoke detection failure in Section 6.1 is presented in Example 6.1 shows a value of 8.41E-inconsistent with Figure 61 and Appendix C, Tables C2 through 02 (revised to 2.6E-02 based CMWRA29) and Chapter 6 C17 equations. Smoke detection ineffectiveness and unreliability results from:
are mutually exclusive so these values should be added together when calculating the first interruptible detection failure Automatic Smoke Detection (OR Gate):
probability. This impacts all examples through Chapter 6 and Unreliability: 0.05 Appendix D Smoke detector ineffectiveness: 0.07 0.05+0.07 - (0.05 x 0.07) = 0.1122 Personnel Detection: 0.231 First Detection AND Gate:
0.231 x 0.1122 = 0.02595 Page 8 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date Notes provided below each table in Appendix C present the calculation for the values. Per the note below Table C-2:
(Probability of No Personnel Present [Table 5-6]) x
((MCR Indication Unreliability [Section 5.3.3.2] +
MCR Operator HEP [Section 5.3.3.3]) - (MCR Indication Unreliability [Section 5.3.3.2] x MCR Operator HEP [Section 5.3.3.3])) x (Smoke Detection Ineffectiveness [Table 5-2] + Smoke Detection Unreliability [NUREG/CR-6850] +
Smoke Detection Unavailability [0.01, assumed])
Note: The value for smoke detector ineffectiveness has been revised as the result of a comment related to the Monte Carlo analysis on smoke detection, so the values may have changed but the calculation method remains the same.
NEI-16 As presented in Figure 6-3, smoke detection 08/13/19 Section 6.1: Second Growing: 3.64E01 unreliability and ineffectiveness are included in the Treatment of automatic smoke detection failure in Figure 62 is detection failure calculation under an OR gate, inconsistent with Figure 61 and Appendix C, Tables C22 through resulting in an addition of the two terms.
Chapter 6 C25 equations. Smoke detection ineffectiveness and unreliability are mutually exclusive so these values should be added together The value for the Second Growing detection step for the second growing detection failure probability. This impacts presented with Example 6.1 shows a value of 3.64E-01 (revised to 1.12E-01 based on all examples through Chapter 6 and Appendix D CNWRA29) and results from:
Page 9 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date Smoke Detection Occurs in Time (AND Gate):
Smoke detector ineffectiveness: 0.07 Flag to confirm smoke detection modeling suggests detection occurs prior to the time to damage: 1.0 0.07 x 1.0 = 0.07 Unreliability: 0.05 Second Detection OR Gate:
0.07 + 0.05 - (0.07 x 0.05) = 0.1122 Notes provided below each table in Appendix C present the calculation for the values. Per the note below Table C-22:
(Smoke Detection Ineffectiveness [Table 5-2] +
Smoke Detection Unreliability [NUREG/CR-6850]
+ Smoke Detection Unavailability [0.01, assumed]
Note: The value for smoke detector ineffectiveness has been revised as the result of a comment related to the Monte Carlo analysis on smoke detection, so the values may have changed but the calculation method remains the same.
NEI-17 68 Figure 66: 0.72 & 0.28 Updated Figure 6-6 8/12/19 Page 10 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date Replace 0.72 and 0.28 with 0.723 and 0.277 respectively to be consistent with Table 32 NEI-18 Figure 66: 0.0324581 Updated Figure 6-6 8/12/19 68 Update 0.0324581 to 3.25E02 to be consistent with other similar values.
NEI-19 The guidance in Section 5.3.3.5 states: 08/07/19 Credit for adjacent spaces may only be taken for a maximum of half of the rating of the source compartment.
Table 61: Note: No additional credit is given for the higher The entry in Table 6-1 takes half the credit of a medium rating (+5/2), even though the adjacent maintenance rating. Additionally, no credit may be taken for an space has a High rating (10).
adjacent occupancy with a lower rating.
69 The text should be revised as:
In the current calculation examples, credit is actually given for the higher maintenance rating of the adjacent compartment. Table 61: Note: Credit for the adjacent space may Suggest to revise the note or calculation accordingly. only be taken for values up to that of the source compartment. Therefore, only a value of 5 - equal to the medium rating of source compartment - is credited. A value of 10 - associated with the High rating - is not allowed. Additionally, no credit may be taken for an adjacent occupancy with a lower rating.
NEI-20 Yes, this is correct. Updated the text in Table 6-1 8/1/19 Table 61: (5+0/2)/10+(5+5/2)/50 = 0.0575 under the Low occupancy and higher 69 The second half of the equation appears to be missing that maintenance ratings in adjacent spaces subtracts the adjacent compartment's higher maintenance rating (Equation 52). It also appears the final value should be updated Page 11 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date to 0.575 with the equation: (5+0/2)/10+(5+5/2)/50 (5+0/2)/10 x (5+5/2)/50 = 0.575 NEI-21 Table 62 Added. 8/12/19 Add "Pns =" in front of the Pns value in the Automatic Smoke 610 & 611 Detection Failure Probability and Pns column. Similarly the detection failure probability should be denoted as well to avoid misinterpretation.
NEI-22 For the purposes of fire modeling, the NUREG/CR6850 growth For interruptible fires the analyst has two options. 8/5/19 profile may be used with a suggested consideration for They can use the specific interruptible fire profile interruptible fires. OR they can use the NUREG/CR-6850 profile with 71 pre-growth. The text in the key findings, Section 4, This statement appears to suggest that the Figure 43 and Section 7 have been updated to make this interruptible curve cannot or should not be used as opposed to clear and consistent.
the old 6850 curve with a delay. Clarification may be warranted.
NEI-23 For interruptible fires, a period of up to 8 minutes with no See revised guidance in Section 4.2. 8/5/19 measurable heat release rate (HRR) may be included prior to the period of fire growth.
71 Suggest adding a statement that use of 4 minutes is recommended for the pregrowth period with reference to section 4.2.
NEI-24 7.5 NSP Estimation Update Added summary of the NSP floor for the MCR in 8/12/19 There is no mention of the revised NSP floor for the MCR in the Section 7.5 73 summary. If this was primarily developed by NUREG2178 V2, this should be noted since the results are still presented in this NUREG as well.
Page 12 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date NEI-25 This was updated to include the new cabinet fires 08/07/19 NUREG2169 l All Fires l 457 l 6691 occurring since 2009. Additionally, per the "All Fires" number of events and total duration do not match discussion in Section 3.5.1, events occurring prior Table 73 NUREG2169, Table 51. If this has been updated by NUREG2230 to 1990 have been excluded from the suppression or NUREG2178 V2, the reference should be updated. rate assessment. The source column entry has been updated to: NUREG-2169, NUREG-2230 NEI-26 The 457 is incorrect. This value comes from an 08/07/19 All Fires l 457 l 6691 older iteration of the developed methodology.
In previous versions of this chart, such as NUREG2169 Table 51 As discussed in Section 3.5.1, suppression times and NUREG/CR6850 Supp1 Table 142, the "All Fires" curve is the for events from the period between 1980 and 1989 Table 73 sum of all the previous line items for events and duration. This were excluded.
does not appear to be the case for NUREG2230. If a differing No change was made in the methodology of how methodology was used, explanation or disclaimer should be the All Fires category is determined. The correct provided. listing should be:
NUREG2169, NUREG2230 l All Fires l 398 l 5878 NEI-27 Added
Reference:
08/22/2019 NUREG-2178, Volume 2/EPRI 3002016052, Refining and Characterizing Heat Release Rates From Electrical REFERENCES Enclosures During Fire (RACHELLE-FIRE), Volume 2:
Fire Modeling Guidance for Electrical Cabinets, Electric 81 If NUREG2178 V2 is referenced in the document, it should be Motors, Indoor Dry Transformers, and the Main Control provided as a reference as well. Board, Draft for Public Comment, U.S. Nuclear Regulatory Commission, Washington, DC, and EPRI, Palo Alto, CA, 2019.
And added cross-reference in text.
EDF-1 Section 1 First, EDF fully agrees with the need to improve the modeling of No changes to text. Comment reflects similar 8/1/19 the fire scenario progression to better reflect Operating project undertaken by EDF.
Page 13 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date Page 11 Experience, especially for electrical cabinets which represent major contributors to the FireCDF. As you did, EDF noted that Fire PRA insights were not consistent with French OPEX which shows that most fires are contained and limited to the ignition fire source thanks to prompt suppression.
Therefore, EDF developed in 2018 its own approach that consists in a deeper analysis of the French OPEX to distinguish various fire progression scenarios. Whereas the philosophy that inspired both yours and our methodology is similar, both methodologies differ.
Comments that follow are directed towards your methodology proposal.
EDF-2 Section 3.3.1.5 they do not continue to growth prior to Corrected 7/31/19 page 37 EDF-3 Section 3.4.1 The document should define the application scope of the so Scope revised as: 08/20/2019 page 38 defined Interruptible Fire and Growing Fire Split Fraction as The scope of the methodology described in this report is regards the quantification tasks of the FPRA methodology. limited to electrical cabinet sources (Bin 15, electrical cabinets) with detailed fire modeling.
For the FPRA Task 7 Quantitative screening, the fire ignition frequencies to use shall not distinguish interruptible / growing fire. Such distinction shall only be made through Task 14 Fire risk quantification and detailed fire scenarios analysis.
EDF-4 Section 4.1.1 It is proposed to define a pregrowth period with zero HRR for Agree, the experimental results are likely more 8/12/19 page 44 interruptible fires to reflect the time period over which prompt severe than real events (for Interruptible and suppression can prevent the fire growth. Growing Fires).
Page 14 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date The 8 min duration that is proposed is derived from experimental Under the proposed classifications if a fire was results. I believe (but, correct me if I am wrong) that experimental found to damage targets outside the fire ignition tests have been conducted with the objective for the fire to source, it is classified as a growing fire since it develop, i.e. involving specific igniters. They should still be more grew large enough to damage targets outside the severe than real events. equipment. No recorded fire event has damaged equipment after suppression activities have Therefore, I believe that associating the 8min pregrowth period started, so there would be no fraction of fires that derived from experimental tests with the fire nonsuppression could be classified as non-interrupted, interruptible probabilities derived from the fire events will still lead to non fires.
realistic results in terms of the proportion of interruptible fires Applying the proposed guidance in detailed fire non suppressed before any damage out of the fire source modeling should not result in any instances where equipment. damage occurs during the pre-growth period, regardless of how the probability of non-Then, I suggest to further use the fire events from the FEDB to suppression determined or used. This is because define a split fraction to distinguish Interrupted and non during that period there should be no HRR Interrupted interruptible fires. The Interrupted interruptible fires modeled, and therefore no critical HRR that could would be defined as interruptible fire events which have been damage equipment (it is essentially a supplemental suppressed before any damages out of the fire source equipment. period of time that may be considered in the The Noninterrupted interruptible fires would be defined as Interruptible fires path analysis).
interruptible fire events which have not been suppressed before Totally removing the fraction of fires does not seem any damages out of the fire source equipment. appropriate as it is unknown that should the For the Interrupted interruptible fires fraction, as by definition the attempts to suppress the fire failed, there would be damages are limited to the fire source equipment, there is no need no risk associated with those events. Instances where the initial suppression failed would be to carry out any fire physical calculation to define the timing of represented by the events that identified multiple damages.
suppression efforts were required and those For the Noninterrupted interruptible fires, fire physical events are classified as growing fires. Totally calculation shall use the HRR profile defined fir interruptible fire removing this time period or fraction of events does without any pregrowth period (meaning the pregrowth period Page 15 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date in Figure 43 shall be suppressed) as this is considered through not match the practiced method currently practiced the Interrupted/Noninterrupted split fraction. In addition, the in Fire PRA applications.
fire suppression probabilities to be applied to NonInterrupted interruptible fire shall be established considering only Non Interrupted interruptible fire events from the FEDB (meaning the fire suppression probabilities curve in Figure 3.3 shall be updated).
CNWRA1 General As the reader is going through the document, it is difficult to The table captions have been updated in resolving 8/30/19 determine and keep track of which event data were used in comment NEI-3 to show the period of data used for the development of each of the elements of the each element presented. For example:
methodology listed in Section 1.1. It would be helpful to The caption associated Table 3-2 presenting the include a summary table (e.g., in an Appendix) that specifies Interruptible and Growing fire split fractions now the period in the event databases that was used for each includes the time period (2000-20140. A similar element. For example, the table would show that event data change is made for the suppression rates (1990-for the period between 2000 and 2014 were used to 2014). The table presenting fire ignition frequency determine the split fractions reported in Table 3-2. The table already includes the time period (2000-2014).
would also explain why the period was chosen. In the example, the reason is that earlier event data was not sufficiently detailed to classify a cabinet fire as interruptible or growing. A reference to the table could be included in Section 1, Introduction.
CNWRA2 Page 3-1, Lines 5-11 There is some repetition in the first paragraph of Section Retained the first few sentences as introductory 8/12/19 3.1, which makes it confusing for the reader. It is suggested paragraph to bridge the gap between Sections 2 to delete the first four sentences in the section, i.e., start and 3. Reduced repetition and revised the final with Following the approach described in NUREG/CR-6850 sentence as suggested.
Alternatively, the four sentences could be a separate introductory paragraph. If these four sentences are retained, Page 16 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date it is suggested to reword the fourth sentence as follows: In other words, these are fires that were not fully involved, did not impact surrounding equipment, or did not damage cable trays or conduit nearby.
CNWRA3 Page 3-8, line 4 Replace Section 3.3 with Section 3.3.1 Updated Section link to 3.3.1 8/12/19 CNWRA4 Page 3-9, Table 3-2 Revise table caption to read Interruptible and growing split Corrected with comment NEI-3 8/12/19 fractions based on 2000-2014 event data.
CNWRA5 Page 3-9, lines 11-12 It is stated that a unique suppression rate associated with Revised as suggested. 08/13/19 the MCR already exists. It is suggested to add a reference to NUREG-2169.
CNWRA6 Page 3-11, lines 8-10 This sentence is confusing. Some minor rewording is Reworded as suggested. 8/12/19 suggested as follows: The interruptible and growing counted under Bin 15, as defined in NUREG/CR-6850, based on events that occurred between 1990 and 2014.
CNWRA7 Page 3-11, lines 10-13 It is stated that 14 of the Bin 15 electrical cabinet events in Yes, this comment is correct. We had enough 08/13/19 this period did not contain enough information to supporting information to classify the 2000-2014 categorize them as interruptible or growing fires. The time period, but not the 1990-1999 period.
preceding sentence seems to imply that this statement Additionally due to the density in the data for the refers to the period between 1990 and 2014. However, it is 2000+ time period the 1990s data is not used to assumed that the period is actually limited to the years calculate frequency (consistent with NUREG-between 1990 and 1999 because between 2000 and 2014 2169).
the event database contained sufficient information to categorize cabinet fires as interruptible or growing. Please confirm or clarify. Revised as:
According to Table A-2, it appears that there were 39 Bin 15 Eight of thirty-nine Bin 15 electrical cabinet events fires in the event database for the period between 1990 and that occurred between 1990 and 1999 did not 1999. Nine events are assigned a fire classification of contain enough information to categorize them as Page 17 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date interruptible fires, while six are considered growing fires. interruptible or growing fires. To capture these Ignoring incident number 20357 for which Pns is events, half of the count and suppression time was undetermined leads to 14 events that occurred between split evenly between the two classifications.
1990 and 1999 and that were included in the analysis to Eighteen of the events occurring in the 1990s determine the additional fire suppression rates. Please could be classified as interruptible and growing and confirm that these events are the 14 events that are referred were considered in their respective suppression to in the aforementioned statement. If so, explain why the rate calculations. For the remaining events, no remaining 25 events in Table A-2 were not included. suppression time could be determined from the In line 12 it is stated that half of the count and suppression event reports.
time was split evenly between the two classifications.
However, Table A-2 indicates that nine of the fourteen Events for the 1990-1999 period are:
cabinet fires were considered interruptible and only five were classified as growing. Please clarify. 12 Interruptible (1 N/A from suppression for self-extinguished, Event 98)
Finally, what are the criteria that were used to classify any of these 14 cabinet fires as interruptible or growing? If the 8 Growing (1 N/A for self-extinguished, Event events were indeed split evenly between the two 20357) classifications, it is worth mentioning that this approach is 12+8=20-2=18 (1990-1999 period events included conservative compared to using the split fractions in Table in in the suppression rate calculations). The 3-2 to estimate how many of the 14 cabinet fires were Interruptible vs. Growing classification is known for interruptible. these events and they are included in the suppression rate calculations similar to those events in the 2000-2014 period.
Of the remaining (39-20= 19) events, the suppression time for 8 were able to be estimated from the information in the event reports. The count and suppression time for these events was split 50/50 between the Interruptible and Growing suppression rate calculations.
Page 18 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date Two of the events were excluded, one event (20302) occurred at location outside the scope of the Fire PRA and the other (20276) self-extinguished.
The suppression time for the remaining 9 events was not able to be determined. Therefore, they could not be used. A note Suppression time indeterminate has been added to the suppression notes for clarification. These events are:
20267, 20268, 20269, 20273, 20275, 20282, 20312, 20334, 20346 CNWRA8 Page 3-12, line 15 Add The latter is usually 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> (see NUREG-2122). at Revised as suggested. 08/13/19 the end of the paragraph. Added NUREG-2122 to reference list.
Page 19 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date CNWRA9 Page 3-12, line 25 The generic frequency for the MCB provided is 2.05E-3 and The applicable modes for each bin originate from 8/22/19 reference is made to NUREG 2178, Volume 2. Section 8.3.1 NUREG/CR-6850 and carried forward in NUREG-of NUREG 2178, Volume 2 shows this generic fire 2169. For Bin 15, this assumption has been carried frequency split by power mode, either Full Power Initiating forward into NUREG-2230, although, the event Event (FPIE) or Low Power Shut Down (LPSD). It seems experience includes one or two LPSD events that appropriate to apply a split fraction to refine the generic may not have been possible if the plant was at frequency, but the floor value analysis only uses the value power. Due to the wide variability of components for FPIE. contained within this bin, this item has been What is justification for not considering LPSD conditions in reserved for future research.
the development of the generic fire frequency for the MCB? For the main control board, upon looking at the experience it was determined that maintenance/work performed in event 51002 would have not occurred during power operations. To reconcile, the fire ignition frequencies for the main control board have been split between at power and LPSD.
Aside from Bin 4 and Bin 15 (AA), the values for Bin 6 and 7 use the at power frequencies (there are different values for LPSD).
CNWRA10 Page 3-12, line 30 Replace (NUREG-2230) with (see Table 3-9) Agree. Replaced NURG-2230 with Table 3-9. 8/12/19 Page 20 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date CNWRA11 Page 3-13, line 2 Provide a basis for the 300-700 range of the number of Revised as: 08/22/2019 cabinets counted as ignition sources in a single unit NPP.
- The electrical cabinet frequency is based on the This seems inconsistent with the average plant-wide count number of cabinets counted in the MCR in the of 750 in Table 6.2.3 of IMC 0308 Attachment 3 Appendix F. FPRA. This count can vary widely from NPP to NPP. The generic frequency value is 3.43E-2/yr (Table 3-9), which applies to a single NPP.
Component counts for both BWRs and PWRs were reviewed and average Bin 15 component count is around 800 per unit. The lowest observed count was over 300. To generalize the analysis, it is assumed that the apportioning factor is a random variable sampled following a uniform distribution with a range between 1/300 and 1/1300. The practical implication of this assumption is that there are on average 800 cabinets counted as ignition sources in a single unit NPP and align with experience performing FPRAs. Since the Bin 15 fire frequency is divided amongst the ignition sources, the lower the component count, the higher the ignition frequency per component. Using a lower bound of 300 provides reasonable assurance that the ignition frequency is overestimated.
CNWRA12 Page 3-13, line 10 Provide a basis for the 10%-30% range of the apportioning Revised as: 08/22/2019 factor (fraction of the control, auxiliary and reactor building * [] To generalize the analysis, it is assumed that floor area representing the MCR). the apportioning factor is a random variable sampled following a uniform distribution with a range between 10% and 30%. The practical implication of this assumption is that, on average, Page 21 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date 20% of the transient frequency is in the MCR. This conservatively assumes that the MCR, often a single room, represents 20% of the combined control, auxiliary, reactor building floor area.
CNWRA13 Page 3-13, lines 19-21 Delete the sentence This results 1.21E-4/day. because Revised as suggested. 08/13/19 and Equation (3-2) this calculation is incorrect, as it does not account for the random variables (total number of cabinets in the plant and apportioning of the area of the MCR). Instead, change Equation 3-2 as follows:
r(24 )=(3.7E-03)/3651E-05 (3-2)
CNWRA14 Page 3-13, line 33 According to NUREG-2178, Volume 1, the 98th percentile At 3 minutes, it is very likely that the fire is still 08/20/2019 peak HRR of a large open electrical enclosure with growing and spreading within the cabinet. At 63kW thermoplastic cable contents is 1,000 kW. Five minutes after it is unlikely that the MCR is abandoned due to ignition the fire in such a cabinet would grow to 174 kW. either heat flux, temperature, or optical density.
Three minutes after ignition the HRR is 63 kW. A value Therefore, 5 minutes will remain as the lower closer to 3 minutes may be a more appropriate lower limit bound.
for the range.
CNWRA15 Page 3-14, line 2 Explain how the non-suppression floor value for a dual unit Table 3-5 entry revised as: 08/13/19 MCR was estimated. Two-unit NPP:
Double the generic frequency associated with the MCB, electrical cabinets, and transients CNWRA16 Page 3-14, line 19 Replace Section 3.5 with Section 3.5.1 Agreed. Updated reference section. 8/12/19 Page 22 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date CNWRA17 Page 3-17, Table 3-9 The frequency for electrical cabinet fires is for AA power The applicable modes for each bin originate from 8/22/19 modes (not separated by AP or AL power modes). NUREG/CR-6850 and carried forward in NUREG-Moreover, the power mode in the table is AA, but the PRA 2169. For Bin 15, this assumption has been carried type is FPIE. This is confusing, since it is explained in forward into NUREG-2230, although, the event Chapter 8 of Volume 2 of NUREG 2178 that the split experience includes one or two LPSD events that between FPIE and LPSD is related to power mode. may not have been possible if the plant was at Please explain why the frequency is split by power modes power. Due to the wide variability of components for the MCB but not for electrical cabinet fires. Clarify the contained within this bin, this item has been difference between PRA types and power modes and how reserved for future research.
that affects the fire frequencies used in the analysis. For the main control board, upon looking at the experience it was determined that maintenance/work performed in event 51002 would have not occurred during power operations. To reconcile, the fire ignition frequencies for the main control board have been split between at power and LPSD.
This approach is consistent with NUREG/CR-6850 where some fire ignition frequency bins are applicable to all modes. There are some fire ignition source bins that may have different factors including maintenance, storage of combustibles, welding, etc. that are not relevant to a mode of operation and the fire ignition frequency is split into at power and LPSD conditions.
CNWRA18 Page 4-3, line 18 and The averages of the pre-growth, growth, steady burning and The purpose of this research is to increase realism 08/30/2019 Table 4-1 decay durations may not be the most representative or in Fire PRA. Using the insights from the FEDB, the appropriate values to use. For example, if the dataset FEDB experience overwhelmingly suggests that consists of a large number of values that are clustered the fire is extinguished before external target together and a relatively small number of values that are damage occurs. Therefore, simply selecting the Page 23 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date much larger, is likely to be the more representative statistic more conservative of the two is not an appropriate to use. It is suggested to add a row with the median values recommendation.
at the bottom of Table 4-1 as shown below, and to use the The mean will be the metric used to estimate the median when that is more representative or conservative. HRR profile period durations:
The mean better captures the range of the distribution, which is important for realism.
The skew in the experimental evidence is also seen in the OPEX, which may not be appropriately reflected in the PRA when using the median.
A short discussion is added in Section 4.2.
Note: A new set of criteria have been established Using the more conservative values of the average or the and applied to the data determined to represent median leads to pre-growth and time-to-peak values of 6 Interruptible fires and presented in Table 4-1. A minutes each, followed by 5 minutes of steady burning and new discussion has been added to describe the a linear decay to a HRR of zero over a period of 13 minutes. criteria and resulting test data used to estimate the pre-growth period duration of an Interruptible fire.
CNWRA19 Page 4-5, line 16 and The previous comment applies to the HRR profile for See response to CNWRA18. 08/30/2019 Table 4-2 growing fires. The additional row with the median values for the data in Table 4-2 is shown below.
Again, using the more conservative values of the average or the median leads to a time to peak of approximately 10 minutes, followed by 9 minutes of steady burning and a linear decay to a HRR of zero over a period of 26 minutes.
Page 24 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date CNWRA20 Page 4-7, lines 7-14 The median time to decay in Table 4-2 is 19 minutes, which See response to CNWRA18. 08/30/2019 is identical to the decay time in NUREG/CR-6850. Using the median decay time of 19 minutes (instead of the mean decay time of 26 minutes) provides a rational basis for softening the effect of long duration fires with relatively low intensity in the later stages of the experiments.
CNWRA21 Page 4-7, lines 24-26 On page 4-7, the final recommended guidance allows the Agree, see the response to comment RC-2 for 8/12/19 and Page 4-8, lines 1-8 analyst to select a time of up to 8 minutes for pre-growth of resolution.
the fire (0 kW). On page 4-8, it is explained why a period of 4 minutes was selected in the examples discussed in Chapter 6.
It is likely that users of this document will use the examples from Chapter 6 as a template and use a 4-minute growth time by default. Consider adding language to explain by example how an appropriate value of the pre-growth time can be estimated for a specific NPP and set of conditions.
Alternatively, if 4 minutes is the recommended pre-growth time, consider modifying the language on page 4-7 to be consistent and provide justification for this recommendation.
CNWRA22 Page 5-3, lines 8-9 Replace The Pns for each profile is summed together to Revised. 8/12/19 determine the total scenario specific NSP. with The total Page 25 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date scenario-specific NSP is determined as the weighted sum based on the split fractions of the Pns for each profile.
CNWRA23 Page 5-6, Table 5-2 The Monte Carlo analysis will have to be redone and the Monte Carlo analysis has been revised. See 08/22/2019 and lines 1-2 resulting probabilities of no detection in Table 5-2 will need individual comments below.
to be updated. Details are provided in the technical comment section on Appendix B.
CNWRA24 Page 6-3, line 5 Start a new paragraph with Because the MCC because Revised as suggested 8/12/19 this sentence starts the discussion of the development of the solution of the NSP event tree according to the new methodology, while the previous sentence ends the discussion of the solution obtained according to the methodology in NUREG/CR-6850 Appendix P.
CNWRA25 General (but related to It would be helpful to include Table 7-3 in the executive Table 7-3 is included to provide a single location 08/22/2019 Table in Section 7) summary. reference for the currently published suppression rates as of the date of the publication of NUREG-2230. While multiple suppression rates are revised as a result of this analysis, the totality of the suppression rates is not appropriate for the executive summary on a methodology focused on a single Bin.
No changes made.
CNWRA26 Table A-2 Explain in the text how cabinet fires that occurred between See response to comment CNWRA7. 08/22/2019 1990 and 1999 were classified as interruptible or growing (see comments on Page 3-11, lines 10-13).
Page 26 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date CNWRA27 Appendix B - General In light of the comments provided below, it appears that the Monte Carlo analysis revised. See responses 08/22/2019 Monte Carlo analysis described in Appendix B will need to below.
be redone. Appendix B in draft NUREG-2230 is reproduced in Attachment A. A proposed revised draft of Appendix B incorporating the comments with track changes turned on can be found in Attachment B. A clean copy of the revised Appendix B is provided in Attachment C. The proposed revision of Appendix B includes some editorial changes that are not discussed here. Comments below and parts of the text that require additional discussion are highlighted in yellow.
CNWRA28 Page B-1, lines 13-14 The purpose of the Monte Carlo analysis is to determine the While the time to activation is not used in the 08/23/2019 probability that a smoke detector will actuate for electrical methodology outlined in this report, it does function cabinet fires. A probability is calculated for each type of as a surrogate of activation that can be judged enclosure that is listed in Table 7-1 of NUREG-2178, against available experimental results, allowing for Volume 1. To determine this probability, it is not necessary a judgment on the validity of the model.
to calculate the time to activation. One only needs to determine for each set of sampled parameters whether the sampled peak HRR is sufficient to result in smoke detector activation. The detector may actuate at a lower HRR, i.e.,
during the growing phase (i.e., the peak HRR is a sufficient but not a necessary condition for detector activation) but this does not affect the count of Monte Carlo realizations that result in detector activation.
CNWRA29 Pages B-1 and B-2, Proposed changes to the table are summarized below. Gamma distribution updated with larger data set. 08/30/2019 Table B-1 Monte Carlo analysis revised to randomly select a cable jacket type from available references and Page 27 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date
- Because it is not necessary to calculate the time of smoke use a soot yield, heat of combustion, and radiative detector activation, the HRR profile duration parameters can fraction for individual iterations. This allows us to be deleted. produce realistic value sets for the analysis.
- Figure B-1 in Section C.2 indicates that a gamma distribution is not a particularly good fit. A better-fitting To address the comments related to the validity of distribution should be used. the SFPE/Tewarson data, we've reviewed the
- Two ranges should be defined for the smoke yield, one for current simulation results. While the average TP and another for TS cables. The former should be used detectable HRRs are in the ranges associated with for fires of cabinets that contain TP cables, the latter for the comment noted reference (Gottuk, Mealy, and cabinets with TS/QTP/SIS cables. The SFPE handbook Floyd) of 50 - 100 kW, the results are heavily
([29]; Table A.39) provides smoke yield data for two types of skewed towards HRRs of 10-30 kW. This range TP cable (PE/PVC and PVC/nylon/PVC-nylon). The values better fits the testing used to develop the range between 0.076 g/g and 0.136 g/g. The range of the SFPE/Tewarson data. Therefore, we believe the smoke yield for TS cables is 0.082-0.175 g/g. Also, some SPFE/Tewarson data is appropriate for the Monte consideration should be given to the fact that the smoke Carlo analysis.
yield values in the SFPE handbook appear to be too high (see Gottuk, D., Mealy, C. and Floyd, J., Smoke Transport Text revised as:
and FDS validation, Proceedings of the Ninth International Where ys is the soot yield of the fuel in kg/kg and IAFSS Symposium, pp. 129-140, 2008.
all other parameters have been identified. The soot doi:10.3801/IAFSS.FSS.9-129). Using smoke yield values yield values used in the Monte Carlo sampling that are too high overestimates the probability of smoke process are taken from the values for electric detector activation.
cables in the SFPE handbook [29; Table A.39].
- The radiative fraction should be varied. A higher radiative The value used for the soot yield - and fraction results in a lower convective fraction and, therefore, subsequently heat of combustion and radiative a lower probability of smoke detector activation. fraction - for each sample is selected randomly Consequently, varying the radiative fraction is important. from the values for electric cables listed in the NUREG-1805 recommends using a generic value of 0.3 but Society of Fire Protection Engineers (SFPE) states that the radiative fraction for different types of fuels handbook [29; Table A.39].
varies between 0.15 and 0.60 (see Chapter 5 in NUREG-Page 28 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date 1805). The low values are for light gaseous fuels, alcohols, Change required updates to:
etc. and are not relevant here. The radiative fraction for Table 5-2: 08/22/2019 cable fires can be estimated from the ratio of the radiative to the chemical heat of combustion reported in Table A.39 of Figure 5-12: 08/22/2019 the SFPE handbook [39]. The resulting range for the Examples (Most figures and tables): 08/22/2019 radiative fraction of TP cables (PE/PVC and Appendix C Tables: 08/23/2019 PVC/nylon/PVC-nylon) is 0.37-0.63 with a mean of 0.50 and Appendix D Tables 08/30/2019 (includes changes median of 0.49. For TS cables (EPR/Hypalon, XLPE/XLPE, to IF supp rate, and propagation through report, and XLPE/neoprene) the range is 0.34-0.62 with a mean of examples, and appendix 0.47 and median of 0.50. Consequently, a single range of 0.35-0.63 for the two cable types seems reasonable.
However, the last part of the previous comment (i.e., the observation that the soot yields in the SFPE handbook are too high) also applies here and indicates that further investigation to find a suitable range for the radiant fraction may be warranted.
CNWRA30 Page B-3, Subsection A sentence is added to indicate that the ceiling jet consists Revised as suggested. 08/22/2019 on Ceiling Jet Density primarily of entrained air and that for estimating its density, the combustion products can be neglected.
CNWRA31 Pages B-3 and B-4, There is no need to introduce this factor. This subsection is Revised as suggested. 08/22/2019 Subsection on Dilution renamed to Normalized Ceiling Jet Mass Flow Rate Factor because that is what is calculated in this subsection.
CNWRA32 Page B-4, Subsection This subsection is renamed to Soot Mass Concentration Revised as suggested. 08/22/2019 on Soot Density because that is what is calculated in this subsection.
CNWRA33 Pages B-5 through B-8, It is proposed that this section be deleted. As mentioned in While the time to activation is not used in the 08/22/2019 Section B.3 the comment on Page B-1, lines 13-14, to determine methodology outlined in this report, it does function Page 29 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date whether a smoke detector will actuate for a set of sampled as a surrogate of activation that can be judged parameters it is not necessary to calculate the time to against available experimental results, allowing for activation. Consequently, an evaluation of the capability of a judgment on the validity of the model.
models to predict smoke detector activation time has little or The average model bias and standard deviation no bearing on the validity of the results of the Monte Carlo are 1.18 and 0.48 respectively.
simulation. Moreover, Figure B-2 indicates that the predictive capability of the two models is very poor. More For comparison the average bias and standard work is needed to investigate the reasons for the deviation for FDTs, CFAST, MAGIC, and FDS discrepancies and to improve the predictive capability of the reviewed in Supplement 1 to NUREG-1824 are models. 1.38 and 0.42, respectively.
This comparison suggests that the model used in the draft NUREG is comparable to the models currently validated for use in Fire PRA.
CNWRA34 General - Appendix D Appendix D does not appear to be referenced anywhere in Text added to Section 5.3: 08/13/19 the main part of the document. Sensitivity to the parameters are analyzed in the examples (see Section 6). Uncertainty in the development of the parameters is reviewed in Appendix D.
CNWRA35 Page D-7, lines 1-2 Based on Table D-1 (page D-4, first two rows), the assumed Text added: 08/13/19 pre-growth time is also a significant sensitivity parameter. Sensitivity to the interruptible fire pre-growth period Please add this parameter to the discussion on page D-7 to duration is also expected since it provides a period recognize its impact. of time where damage is not modeled to occur for nearly 70% of the fires in electrical cabinets.
CNWRA36 Abstract Replace data has been with data have been Revised sentence to read: 8/12/19 Page v, line 6 because data is plural. As a result, further development of the methods has been performed and additional data have been collected and analyzed over the past decade.
These improvements have allowed gaps in the Page 30 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date methods to be closed and more realistic estimates of risks to be obtained.
CNWRA37 Page 2-1, line 26 Replace NUREG/CR-7197 with NUREG/CR- This report is now consistently referred to as 8/12/19 7197 (HELEN-FIRE) [11] because it is referred to as NUREG/CR-7197.
HELEN-FIRE in other parts of the document (e.g., captions to Figures 4-1 and 4-2).
CNWRA38 Figures 2-1 through 2-5 It would be helpful to increase the size of the colored Sizes of legends increased. 08/22/2019 squares in the legends to the pie charts and used different If any of the percentages presented in the Figures hatched patterns so that it easier to distinguish the different contained in Chapter 2 are directly influential in the categories in a B&W hardcopy. methodology presented in later chapters, all necessary discussion is repeated. No confusion as a result of a black and white copy is expected.
No changes made.
CNWRA39 Page 3-11, line 13 Replace Due to the limited event information with Due Revised as suggested. 8/12/19 to the limited cabinet fire event information CNWRA40 Page 3-12, Table 3-4 Please spell out the column headings in this table, i.e., SS = Revised as suggested. 08/13/19 Sum of Squares, df = Degrees of Freedom, etc.
CNWRA41 Page 3-13, lines 7-8 There appears to be a typo in NUREG-2169. The location Current plans are to revise NUREG-2169 within a 08/20/2019 for Bin 7 in Table 4-4 is described as the Diesel Generator year. Correction will be addressed at that time.
Room. Based on Table 6-1 in NUREG/CR-6850 this should be Control/Aux/Reactor Building. It is suggested adding a footnote to point this out. No changes made.
CNWRA42 Page 3-13, lines 33-34 Replace occurs before a half of the growing time with Text re-written / comment no longer applicable. 8/12/19 occurs before half of the growing time Page 31 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date CNWRA43 Page 3-15, Figure 3-4 Growth Fire is referred to as Growing Fire in other Legend revised to Growing. 08/13/19 places. y-axis revised to scientific notation.
Strictly speaking, the lower limit for the y-axis is 0.00001.
CNWRA44 Page 4-1, line 22 Replace Valtion Teknillinen Tutkimuskeskus with VTT is initially identified by its native name in 8/13/19 Technical Research Centre of Finland NUREG-2178, Vol. 1 - the most recent document published reviewing electrical cabinet fire HRRs.
Therefore, the same will be done so here.
No changes made.
CNWRA45 Page 4-1, line 24 Replace NUREG/CR-7197 [11]. with NUREG/CR- Report now consistently referred to as 8/12/19 7197 (HELEN-FIRE) [11]. NUREG/CR-7197.
CNWRA46 Page 4-2, line 12 Replace (Figure 4.2.A) with (e.g., Figure 4.2.A) Revised as suggested. 8/12/19 CNWRA47 Page 4-2, line 13 Replace (Figure 4.2.B) with (e.g., Figure 4.2.B) Revised as suggested. 8/12/19 CNWRA48 Page 4-3, line 19 Replace Chesapeake Bay Detachment (CBD) with CBD is how the tests are characterized in the 08/23/2019 (NIST) because CBD is essentially unknown and the reference, NUREG/CR-7197.
tests were conducted by NIST. No changes made.
CNWRA49 Page 4-5, line 17 Replace Chesapeake Bay Detachment (CBD) with CBD is how the tests are characterized in the 08/23/2019 (NIST) reference, NUREG/CR-7197.
No changes made.
CNWRA50 Page 5-3, line 10 Replace are conceptually similarly with are Revised as suggested. 8/12/19 conceptually similar Page 32 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date CNWRA51 Page 5-27, lines 29-30 Replace an motor control center (MCC) with a Revised as suggested (found in two places) 8/12/19 motor control center (MCC) or with an MCC.
EK-1 General The report fails to address the effects of decisiondependent 8/30/19 The objective of NUREG-2230 is to characterize uncertainty on its dataset. The consequence of this oversight is (Specific comments) fire ignition frequency, fire growth timing, and plant that the dataset is invalid as of the date of publication.
personnel response to fire. Fire ignition frequency and fire growth characteristics are independent of Details on explanation of comments:
operator or plant personnel actions (with the The fire statistics developed in the DRAFT NUREG are not helpful exception of hotwork scenarios, and possibly to investigators interested in avoiding catastrophe (core damage testing and maintenance, etc.).
or more importantly, radiation release) because (at a minimum) they are disconnected from protective equipment availability.
This report is to be used in the risk-informed Only when sufficient protective system equipment are approach currently available to licenses under unavailable will catastrophe follow; a relevant example would be 50.48(c). It is meant to be used in the integrated the sustained loss of electrical power in Generation II light water context of leveraging risk insights that account for reactors due to fire. The NUREG should be rewritten to produce the likelihood, consequence, and impact of fire statistics on data related to protective system equipment unavailability due to fire. Such data would be in the form of event scenarios; according to the PRA policy statement occurrence (date, time, where) and consequential protective of 1995. In this context the data set is valid.
system equipment unavailability (list of equipment affected, duration of unavailability); these data would be useful to reliability engineering investigators.
General The report claims the reported data can be used for model Specific to the model validation in the DRAFT 8/30/19 validation but fails to explain how the data can validate a model NUREG, validation of the smoke detection model (Specific comments)
(nuclear plant risk model) in light of the fact that the dataset described in Appendix B is performed following the lacks the necessary correlate (that is, core damage or radiation fire modeling validation methodology in NUREG-release.) 1934. This is the only model validation effort performed in NUREG-2230.
Details on explanation of comments:
Page 33 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date When using empirical data as a predictor of future performance, as proposed in the DRAFT NUREG, there are at least two The data used in this report is relevant and important cautions (enumerated below) that must be observed. appropriate. As noted in the report, the majority of Nowhere in the DRAFT NUREG is it explained how these cautions fires are potentially challenging, and by definition, are being considered. do not damage external targets. Furthermore, the protective system availability is independent of the (a) The data must be meaningful to the observation of interest - fire initiation and treated through the PRA model Lacking statistical data on catastrophe following fire or and/or non-suppression probabilities. Due to the equipment unavailability is the only statistic with relevance to number of components, locations, and cable reliability. Note that the DRAFT NUREG has no information about routing, even if this information was available, protective system unavailability as part of the statistics could not be applied practically.
developed. If the data are not meaningful in this regard, the statistics will not be helpful to investigators who intend to reduce the probability of protective system equipment unavailability due The output of the PRA (both numerically and to fire. Statistics for any complicated process with risk for harm insights) can and should be used to manage and are obtained to understand frequency of harmful events. For improve plant risk (including fire). The suggestions example, data on pedestrian crossing deaths at an automobile to improve availability of protective system and intersection are collected by noting when a death occurs. At decreasing the occurrence of fires are examples of some level of loss of life, actions are taken to reduce the risk changes that can be realized with a Fire PRA.
(traffic deaths) using, for example traffic signals, overhead However, that is not the objective of the walkways, and other means; the data are not collected at a level development of guidance for performing a fire that checks for frequency of human error on braking the PRA. The objective of PRA method development is automobile, driver attentiveness, condition of the automobile, to develop practical, credible, and generically likely speed of approach, weather conditions, etc. Only the applicable methods and data that can be applied in frequency of pedestrian deaths is required to bring about risk a plant-specific fire PRA to extract the insights management action. mentioned in the comment.
Page 34 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date (b) The end use of the statistics is (rationally should be) to The second part of the comment cautions about improve the availability of critical protective system equipment the use of past data for predicting future or decrease the occurrence of fires affecting critical protective performance. The commenter suggests that if the equipment availability (that is, risk reduction). A direct root cause is corrected, then past performance consequence of actions taken to avoid future occurrence may not be representative. The use of OPEX is (depending on the efficacy of the used for calculating initiating frequencies not only actions) invalidates the data used to predict" the future. within the fire domain but also for internal events Prediction of future performance is the domain of decision (reactor trip, turbine trip, LOOP) and internal dependent uncertainty; decisiondependent uncertainty is flooding. Furthermore, the fire ignition frequencies avoided out of necessity by investigators using historical datasets are meant to be updated every few years to by being careful to preserve prior data validity (ensure the future capture changes in performance. Changes in is predicted by the past). For example, hedge fund managers use performance have been observed over the course historical data to develop high frequency trade algorithms and of the fire PRA development. Since the use of past many are very successful. However, unlike risk management in data is an accepted practice in the PRA community commercial nuclear power where workers intend to change the no changes are implemented.
future, the clever hedge fund manager is very careful to ensure her trading volume doesn't affect the past data: she must avoid the temptation to become too greedy".
General The data should be made useful by adding information related to The documentation of the fire events is typically 8/30/19 the efficacy of protection. For example by correlating protective commensurate with the severity of the fire event.
(Specific comments) system unavailability to the fire data, the statistics would be Fires that have minimal damage, minimal useful to investigators. suppression, and/or minimal plant impact are often brief write-ups of the event. More severe events, The prior comments are not entirely naive; the reviewer such as high-energy arcing faults (HEAFs) understands the end use of the DRAFT NUREG statistics is PRA. A generally have more detailed information on the critical point is PRA may or may not contain (have as a basic timeline, plant response, and in some cases an event) the most interesting aspect of a particular fire event that apparent or root cause. In summary, the level of detail suggested is not available for most Page 35 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date bears on risk management the failure mode (or mechanism) potentially challenging fires nor does it invalidate which is the root cause. the current approach or contents of the DRAFT NUREG.
For example, dust buildup in a high voltage enclosure has been shown (this has happened) to result in switchgear fires; a The PRA is not intended to go into the details of particular PRA is very likely to have an initiating event of fire in dust building in a high voltage enclosure. As such, such an enclosure as a basic event. Depending on the enclosure, the NUREG was written to capture details from the the PRA may include a sequence that contains the event as a FEDB to the extent possible, albeit, not at the level cause for protective equipment unavailability. The advantage of that the commenter is suggesting. Such level of PRA over other forms of reliability analysis is that it generally detail would cause the complexity of any approach goes further than FMEA for example by attempting to to increase significantly, well-beyond the value of comprehensively include all scenarios that could be envisioned the current PRA approaches within the context of from the event.
the PRA Policy Statement.
In contradiction to what is suggested in the DRAFT NUREG, supplying more detailed statistics to a PRA in pursuit of The statistics developed and presented in the quantification will not help plant workers or regulators develop DRAFT NUREG are already intended to assist new strategies, different inspections, or tests that would reduce analysts in determining if and what new strategies, the probability of future catastrophe (this point is closely related different inspections, or tests are necessary to to Hansson's \Tuxedo Fallacy"). reduce the risk of core damage or radiation release resulting from fires in electrical cabinets. In Unless this reviewer is missing a more subtle point, it seems the addition, as stated above, the continuous updating objective of the work documented in the DRAFT NUREG is and collecting of information is focused on (should be) intended to help risk management. Such an objective ensuring the risk management of such contributors can only come from engineering analysis of the event, correctly supports decision-making.
understanding the root cause, and developing strategies that would reduce protective system equipment unavailability (in this case from fire). Once actions are taken, the prior data and Use of prior data is only done when the representativeness of the data and time period Page 36 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date concomitant statistics presented in the DRAFT NUREG are considered is assessed. This is not anathema with invalidated. risk assessment or use of PRA within a risk-informed decision-making process. Otherwise, no prior data could ever be used and no risk management approach would ever be deemed practical.
General Effective risk management requires understanding and The overarching comment seeks to derive the root 8/30/19 addressing the root cause of events that can lead to catastrophe; causes of fires in nuclear power plants with the (Summary of a useful fire database would explain the root cause for fires, objective of reducing the likelihood of fires. While Explanation of Comments) possibly categorizing similar causes The data in the report should this is a noble objective, this is not the purpose of at a minimum be recast to indicate events where availability of NUREG-2230. Nor is this the purpose of the EPRI protective system equipment is affected (reduced). Fire Events Database. The purpose of the FEDB is to gather data on fire events, extract data for input The DRAFT NUREG in a few places suggests the data can be used to fire PRA, and use the insights to inform fire PRA to produce more realistic modeling of fire risk". This claim is methods (e.g., NUREG-2230).
misleading for many reasons but particularly when used in a The primary objective of NUREG-2230 is to setting of dynamic risk management.
enhance the modeling of electrical cabinets in both Models of fire hazard (for example, probabilistic models) cannot the fire growth and the plant personnel response.
be verified using data as proposed in the DRAFT NUREG; for the The draft NUREG presents a methodology model to be verified, data for fire terminating in catastrophe (for example, core damage or radiation release) would be needed. If developed from detailed reviews of operational it were possible to have such data then a model (PRA or some experience from the U.S. nuclear industry, statistical method) could be validated; otherwise, in the absent specifically those associated with fires in electrical ouch data, there is no way to validate nearness" to reality. In cabinets. The result is a method of estimating the point of fact, lack of evidence may be an indication that, while frequency of and the rate of suppression for fires in occurrence of fire is certainly to be avoided, fire is of less concern electrical cabinets that more closely resembles the compared to other protective system breakdowns that have lower rates of risk significant fires documented in actually led to catastrophe. fire event reports when compared to the results of Page 37 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date Dynamic risk management in learning organizations will cause the process recommended to date. Compared to observed fire root causes to be eliminated or significantly the methods and data provided in NUREG/CR-mitigated as they occur; that is, the intent of risk management in 6850, this is an improvement.
learning organizations is change the future based on The appropriateness of the data can be compared observations. This effect, changing operation, maintenance, or to the very limited number of severe fire events design based on root causes in response to fire events can be observed. In addition, verifying more severe fires characterized as decisiondependent uncertainty. with core damage events is impractical since these seldom occur (no approach based on such a concept, probabilistic or deterministic) would provide value for actual risk management. Instead, the current approach is to characterize the distinction between challenging and non-challenging fires, consider how protective equipment may fail while responding to a fire, and evaluate whether the current fire protection posture provides reasonable assurance of adequate protection.
Finally, the comment appears to state that organizational factors need to be considered in risk management. While those aspects are connected to the causes of potential fires, available data and modeling approaches that explicitly include such factors are immature and not currently part of the state of practice.
Page 38 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date Discussion The commercial nuclear power industry is experiencing difficulty The development and deployment of fire PRAs has 8/30/19 competing in the current electrical market leading to the extent led to numerous safety improvements, most that several plants are forced into early retirement; and notably through the implementation of NFPA 805.
therefore plant operators (especially merchants" in deregulated markets) are looking for maximum benefit for resources expended. The methodology in NUREG-2230 was intentionally developed to be adaptable within the This reviewer notes that developing and maintaining a Fire PRA" existing framework of NUREG/CR-6850. NUREG-costs plants millions of dollars. Such large costs would be justified 2230 can be incorporated during the development if they led to, or could be shown to lead to, significant of a new fire PRA, or can be incorporated in an improvements to safety. The comments made in the previous are existing fire PRA methodology. NUREG-2230 to say that expenditures that would add complexity to PRA will requires no walkdowns/target collection and not improve the efficacy of protection against core damage or minimal additional inputs (e.g., if taking credit for radiation release. As stated above, based on basic risk monitored equipment in the MCR). The authors of management practice, this reviewer claims it can not be shown NUREG-2230 are not aware of other statistical that spending resources that would apply the statistics developed analyses that discredit the current use of data or in the DRAFT NUREG the methodology presented, and it is unclear how (that is, add complexity to a PRA) will change the risk of core unmentioned other statistical analyses would damage or radiation release in a nuclear power power plant; perform in terms of cost to industry, complexity, although the assessed values (numbers) may change, it would and technical bases.
have no effect on safety.
This reviewer recommends a more wellfounded statistical This methodology may change the calculated plant analysis be conducted that at least gives estimates of the current fire risk and provide insights that are more state of affairs in terms of fires on protective system equipment consistent with the observed OPEX. The availability. Of course any fire should be prevented if for no other commenter is correct that this method will not reason than worker protection; root change the actual plant risk, as that is only achievable through changes to plant procedures, plant modifications, or changes to other plant Page 39 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date cause analysis and risk management against recurrence practices. The insights from this report can be used (addressing the root cause) is the best path to prevention. by analysts to better focus resources where the greatest improvements may be made. To ensure these actions taken to reduce risk are those that are capable of reducing the true risk associated with core damage or radiation release, a certain amount of complexity may be necessary.
The methodology in NUREG-2230 is an approach based on statistical analysis. The development of the additional detail was necessary based on original methodology oversimplifications. This report describes more distinct treatment for fire growth and suppression and is independent of the consequences (e.g., impact and availability of protective system equipment). Lastly, it is not an objective of this work to reduce the frequency of fires in nuclear power plants. This objective of the fire ignition frequencies is to objectively characterize the occurrence. The plant fire protection group, preventative maintenance, and other industry groups may provide recommendations for reducing the occurrence of fires, however this is not the subject of this report.
From NUREG-2178 Vol. 2 Comment Period Page 40 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date NUREG-2178 Chapter 7 is very confusing as to what the conclusion is providing. By reading it carefully, the chapter seems to propose a floor value for NSP of This comment is in relation to the earlier draft that 1.0E-05. However, Chapter 8 provides a DH-1 7 was posted. This issue has been reconciled in the 8/30/19 recommended floor value of 1E-04 (Table 807). So new version.
the chapters disagree. Please re-write the section to be less confusing and ensure Chapter 7 and 8 agree with the approach and the recommended floor value.
There is no double counting the frequency. The This section is intended to re-evaluate the floor value of the value developed in this section is a 'floor' or limit.
MCR non-suppression probability as a replacement of the This value will serve as the lowest value for which 0.001 value that is currently used. It should be noted that a Pns value would be considered appropriate and Pg 7-1 Line 12-14 GG-1 "the probability of having a fire at some point" in the MCR is acceptable for use in the Fire PRA. It is not used to 9/3/19 Pg 7-2 Line 35-37 already covered by the fire ignition frequency. Therefore, determine the scenario specific Pns or subsequent including it in the non-suppression probability will result in scenario frequency.
double counting the frequency component. The introduction discussion has be simplified to remove references to the Pns model.
While FAQ 12-0064 does allow for a zero (0) Hot Work influencing factor rating where activities during power operation are precluded by []
For at power operation, would not consider Bin 6 transient operation, it also includes the Extremely Low GG-2 Pg 7-1 Line 23 fires due to welding and cutting the MCR to be one of the rating for application to qualified MCRs. Therefore, 9/3/19 typical ignition sources. while unlikely, it is possible Bin 6 transient fires may need to be considered in a MCR analysis.
A reference to the FAQ has been added.
Page 41 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date The count is reflected in the ignition source weighting factor. This apportioning among cabinets within the plant is captured in the uniform This paragraph starts with the discussion the Bin 15 distribution assuming the number of cabinets electrical cabinet frequency is based on the number of ranges from 300 to 1300 cabinets in the entire cabinets counted in the MCR. The calculation then plant as described in the text. The development of GG-3 Pg 7-1 Line 31 seemingly applies the full Bin 15 generic frequency to the the proposed floor value includes a single cabinet, 9/3/19 MCR, whilst in reality the MCR would have a small fraction weighted by the varying count of total cabinets of the total plant wide Bin 15 count and this should be within the plant.
reflected in the calculation.
Discussion was revised to describe that the ignition source weighting factor is credited in the calculation, not the simply the generic frequency.
Saying the MCR represents 20% of the total floor area of A sensitivity case has been added to Table 3-5 GG-4 Pg 7-2 Line 6 the control / auxiliary / reactor buildings is overly assuming the floor area ranges between 1% and 9/3/19 conservative. 10%.
The commenter is correct that the mission time plays no direct role in the calculation of a MCR fire The statement that "the Fire PRA estimates the probability scenario frequency. It does, however, provide a of an event occurring over a 24-hour mission time" does not period duration typically used in PRA analyses.
make sense in the context of fire scenario frequency As noted in the draft NUREG, the goal is to GG-5 Pg 7-2 Line 11-14 calculation. The plant mission time in PRA refers to the time estimate a floor value that represents the best 9/13/19 needed for the mitigation systems to bring the plant in a safe possible manual PNS. This is done considering both long-term condition. Therefore, the mission time plays no the probability of having a fire and failing to role in the calculation of the MCR fire scenario frequencies. suppress that fire.
This consideration is developed under the idea that an unsuppressed Control Room fire within a period Page 42 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date of hours is very low likelihood. This is consistent with the OPEX, which so far suggests fire as suppressed in less than 10 minutes given the continuous pretense of operators in the room.
Since the floor is a value set up to prevent unrealistically low estimations of the PNS, a value based on a 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> period of time sets up a practical limit (i.e. this floor help to prevent a PNS from being unrealistically low as it will be higher than the probability of an unsuppressed fire).
Considering an extreme case, should the probability of having a fire be 1.0, a floor value of 1.0 would not match the OPEX as there is no or is there ever expected to be a fire that is not eventually suppressed. Therefore, some consideration of the probability of suppressing the fire must be included. This is the basis for the dual consideration of both the probability of seeing a fire and failing to suppress the fire.
Considering the generic fire ignition frequencies, the development of a probability of having a fire may be obtained if some decision on an appropriate period duration is made.
Understanding that the floor value exists simply to truncate incredibly low PNS values and does not represent any real value, the period duration is Page 43 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date open to a wide range of interpretation. Some possible durations include:
1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> (a common detailed fire modeling scenario period: 1.0E-08 floor) 217 minutes (longest suppression time in OPEX: 3.6-E08 floor) 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> (assumed MCR operator shift duration: 2.0E-08 floor) 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> (mission time: 2.4E-07 floor) 1 Week (1.7E-06 floor) 1 Month (7.3E-06 floor) 1 Year (8.6E-05 floor) 18 Months (re-fueling cycle: 1.3E-04 floor)
Justifications could be made for each of these proposed, if not alternate durations. The 24-hour period selected as it provides a generic basis for a period duration representative of probabilistic analyses.
No changes made.
The calculation of the non-suppression probability floor value presented in Section 7.3 does not appear to have a The comment is incorrect that the proposed floor in sound basis. If applied in practice as a direct the draft NUREG would result in a scenario GG-6 Pg 7-2 Line 35-37 replacement of the 0.001 floor NSP, this approach would frequency equation similar to: Ignition Frequency 8/30/19 result in the following scenario frequency equation : x...x (Ignition Frequency expressed as a probability Scenario Frequency = Ignition Frequency x...x (Ignition for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />) x (Average NSP).
Frequency expressed as a Page 44 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date probability for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />) x (Average NSP) The only effect on a scenario frequency the floor Where the Ignition frequency appears twice - first directly value would have on a scenario frequency is the and then again as a probability of a fire within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. The point at which the value would be truncated.
double counting of the frequency drives the resulting Floor value discussion revised to clarify that the number unrealistically low.
development of the stated floor value is a calculation of a scenario PNS value.
Eighteen minutes was selected because it is the time it takes to reach the floor value of 1E-3 described in NUREG/CR-6850. Since the new floor would be reached at durations longer than the I would also agree that after an initial period the MCR fire evidence observed in the OPEX, the lower floor NSP could be considered no different than ordinary ignition may not be appropriate as would not be captured source NCP. However, it doesn't seem very clear how the in the data used for developing the manual 18 min. initial period can be justified. The fact that the NSP suppression rate constant.
GG-7 (Email) Email, 09/03/2019 9/6/19 curve reaches 1E-3 at that point doesn't seem to me a To resolve this, the two-step method was strong enough engineering basis. I think the text needs to proposed, with the 1E-3 floor value used as the answer the question why not switch to the other suppression transition point between using the main control curve earlier or later than the proposed 18 min.?
room suppression rate constant and an ignition source specific suppression rate constant that would capture longer duration fires. Given the use of 1E-3, the transition point occurs at 18 minutes.
The first step uses the MCR suppression rate in Table 8-6 Per the text in Section 3.5.2:
for calculations of the Pns with a floor value of 1E-03. This The second step captures all remaining MCR fire 7.4 results in fires that are suppressed prior to approximately 18 NEI-51 durations up to the floor of 2.4E-07 through the use 8/30/19 (7-3) minutes using a suppression rate of 0.385. The second step of an ignition source bin specific suppression rate captures all remaining MCR fire durations up to the proposed floor of 2.4E-07 by making use of the ignition (interruptible, growing, transient, etc.).
Page 45 of 46
REVIEW / COMMENT DOCUMENTATION Reviewer: Various Phone: Document #, Rev: NUREG-2230 / EPRI 3002016051 Methodology for Modeling Fire Growth and Suppression Response for Electrical Cabinet Fires in Nuclear Power Discipline/Department: Various Date: July 2019
Title:
Plants
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT Comments shall be:
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /
Section / Paragraph Date source bin specific suppression rate. For example, a fire in a No changes made.
cabinet located within the MCR will use the Interruptible and Growing suppression rates presented in NUREG-2230 [62]
for fire durations in excess of 18 minutes.
Specify that the two-step method is to be used for sources in the MCR besides the main control board The first step uses the MCR suppression rate in Table 8-6 for calculations of the Pns with a floor value of 1E-03. This results in fires that are suppressed prior to approximately 18 minutes using a suppression rate of 0.385. The second step captures all remaining MCR fire durations up to the proposed floor of 2.4E-07 by making use of the ignition Equations for the calculation of the numerical Section 7.4, Figure 7-1, source bin specific suppression rate. For example, a fire in a suppression results for electrical cabinet NEI-67 and Table 7-2 cabinet located within the MCR will use the Interruptible and 9/6/19 (interruptible and growing) and transient ignition (7-3, 7-4, and 7-5) Growing suppression rates presented in NUREG-2230 [62]
sources have been added.
for fire durations in excess of 18 minutes.
Provide equations so the values in Table 7-2 can be replicated. Statement is unclear how to apply the two-step calculation Page 46 of 46