NUREG-2233, EPRI 3002016054 Comments Resolution
| ML20083F972 | |
| Person / Time | |
|---|---|
| Issue date: | 03/31/2020 |
| From: | Office of Nuclear Regulatory Research |
| To: | |
| David Stroup | |
| References | |
| EPRI 3002016054, NUREG-2233 | |
| Download: ML20083F972 (14) | |
Text
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 1 of 14 Comment No.
Document Number Section / Paragraph Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Reviewer Acceptance /
Date CNWRA CNWRA-1 Eq 2-1 Although the definition of IS in Equation 2-1 is consistent with the definition on page 6-2 of NUREG/CR-6850, Vol. 2, the use of the term plant level creates some confusion with fire ignition frequencies that are mapped to the plant-wide components location. We suggest replacing where IS is the plant level fire ignition frequency for ignition source IS (e.g., values from NUREG-2169 [9]), WL is with where IS is the fire ignition frequency for ignition source IS (e.g.,
values from NUREG-2169 [9]) and location L, WL is In addition, FAQ 12-0064 (discussed in the next section) clarified that the weighting of the influence factors should be done within each location set. We therefore suggest pointing out here that transient fire ignition frequencies are mapped to four (general transients) or three (transients due to hot work) location bins Accept the edit Accept the edit CNWRA-2 2.1.4 Replace Plant Analysis Unit with Physical Analysis Unit.
Accept edit CNWRA-3 Eq 2-8 Replace k with IS,k.
Accept edit
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 2 of 14 CNWRA-4 2-2
- 1.
There appears to be some subjectivity in the decision to accept or reject a specific test. For example, tests that involved more than one bag were rejected because there are no records in the FEDB for fires that involve multiple trash bags.
However, a fire involving two small trash bags may be smaller than a fire involving a single large trash bag.
- 2.
The third test from the top on page 2-8 was rejected because part of the fuel package consisted of PE coated milk cartons, which are not an expected item for a trash can in an NPP. However, the third test from the bottom on page 2-7 involves the same fuel types (PE and paper) and was included.
- 3.
The last three tests on page 2-10 involved a wood crib. A wood crib would be representative of a small pallet fire. Appendix G in NUREG-1934 illustrates the use of fire models to simulate the effects of a much larger pallet fire. Also, before gas burners were specified in room/corner tests, wood cribs were used as a surrogate for a trash can.
- 4.
How did the authors ensure that the more intense transient fires that can occur in NPPs were not underrepresented in the final set of tests that were used?
- 5.
Finally, what is meant by extreme? It would be helpful to include a brief description of the criteria that were used to describe a fuel package as extreme.
- 1.
There is always subjectivity in this sort of evaluation; it is unavoidable. However, the size of the fire is not the relevant metric. What is relevant is the plausibility of the fuel package being a transient fire event in a nuclear power plant. It was deemed plausible that there could be a large or a small bag of trash involved in a transient incident. There are both large and small trash cans present. However, it was not deemed plausible for a collection of bags to present. These arent seen and OE and one would only expect the possibility of multiple bags during housekeeping activities where bags would be promptly removed and very unlikely to see an ignition event in that time.
Changes made to Section 2.2 tables to make this more clear.
- 2.
The PE materials in the accepted tests on page 2-7 were PE trash bag and a single PE wash bottle. These are not the same fuel source as a bin filled with a low melting point (wax) PE coated cardboard.
- 3.
A wood crib is not representative of a pallet fire. Wood cribs are a cubical array that maximize ventilation and heat retention. This is a significantly different configuration than a wood pallet which is a large plat flat plane. Configuration matters as it changes the shape of the heat release curve which impacts the
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 3 of 14 Comment No.
Document Number Section / Paragraph Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Reviewer Acceptance /
Date ZOIs. Furthermore, wood cribs in the referenced tests were ignited with a pan of JP4 (covering half the area beneath the crib) to ensure rapid, uniform ignition of the entire crib. This results in far quicker growth rates than one would get from a wood pallet ignited by typical transient ignition sources (if one could even get the pallet to ignite in the first place).
- 4.
The working group examined every fire event in the FEDB and ensured that fuel packages were present to represent the most severe items seen in OE. This was discussed in the test report.
- 5.
The extreme was in contrast to tests involving multiple gallons of flammable liquid or those tests involving unreasonably large amounts of fuel. Will replace extreme with other wording.
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 4 of 14 Comment No.
Document Number Section / Paragraph Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Reviewer Acceptance /
Date CNWRA-5 3-1 bullet 2.2 We suggest clarifying that TS damage thresholds can be assumed for Kerite FR-II, FR-III, and HT cable targets. FAQ 08-0053, Revision 1 recommends using damage thresholds from NUREG/CR-7102 for Kerite FR-II, FR-III, and HT cable targets. Consequently, assuming TS damage thresholds for Kerite FR-II, FR-III, and HT cable targets is conservative, because the lowest failure temperature reported in Table 8-3 of NUREG/CR-7102 for the Penlight tests performed on these Kerite cable varieties is 367°C (693°F).
Will note that TS could be used for Kerite FR-II, FR-III, and HT cables as conservative values as a means of simplifying application of this guidance for those plants also evaluating TS cables.
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 5 of 14 CNWRA-6 3.2 Is four votes with three persons from the same organization adequate to avoid bias? Why was the vote from the project working group not given more weight?
The three individuals have different educational backgrounds, have walked down different plants for different utilities (they have seen the outcomes of different fire protection programs and oversight),
and work in different offices. There is no reason to think that the simple fact they all work for Jensen Hughes somehow overcomes the diversity of background and experience. Additionally, the working group reviewed the voting process prior to it use and those members were diverse in employer.
The primary author of the test and PRA report developed the initial FEDB partitioning and assignment/weighting of test items.
For the assigning of FEDB events to groups, the end results were presented at a work group meeting and the list of events with assignments was quickly reviewed. No objections were noted by the group. This portion is essentially just one persons effort and giving it more weight is not warranted.
Similar to the FEDB events, the assignment of test items to fuel categories and in category weighting was first developed by the primary author and then presented to the panel. There was more discussion on these and some changes made as a result. This was primarily due to the comments of a couple of members of the group. Again, this was largely the effort of an individual.
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 6 of 14 Comment No.
Document Number Section / Paragraph Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Reviewer Acceptance /
Date CNWRA-7 Page 3-3 Use relative standard deviation instead of relative deviation?
Accept comment CNWRA-8 3-18 Replace Below 10 kW, accuracy, and with Below 10 kW, the ignition source often provides a significant contribution to the HRR measured in the test, the relative errors in yields and the heat of combustion are generally larger due to measurement noise and load cell measurement uncertainty, and Accept comment CNWRA-9 Bullets after table 4-1 Replace a 123 MJ with a 278 kW because the peak HRR is the primary cause of the creation of a damaging gas layer.
Peak HRR is not the primary cause for hot gas layer formation; it is a contributing factor. A 1 MW fire for 1 s isnt going to be as hazardous as a 500 kW fire for 10 minutes. Will edit to note the combination of HRR and TER does not form a layer CNWRA-10 4.1.1 para 2 It does not look like the TER is one third less.
98th TER is 123 MJ. 123 /3 = 40. 123 - 40 = 83.
The 75th TER is 81.5 MJ. This is one third less.
CNWRA-11 4.1.1 The range for TP (20 - 48 cm) seems quite significant 36 cm +/- 12 cm. While the percentage variance is large, the real-world distance is not large which was the comment made. In risk space that spread in distance is not going to have a significant impact. The floor area a fire could occupy and damage a vertical riser is going to be very small for typical compartment sizes across that entire range.
CNWRA-12 5.2.1 Please define Q* and mention that Q* is also referred to as the fire Froude number Comment accepted
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 7 of 14 Comment No.
Document Number Section / Paragraph Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Reviewer Acceptance /
Date CNWRA-13 5.2.1 Many of the NFPA 805 plants used a fire Froude number of 1 across the board. How would the use of 0.54 have affected the ZOI calculations?
All else remaining unchanged in a calculation, a higher Froude number makes the vertical ZOI worse (higher flame/plume height) and horizontal ZOI better (less favorable view factor).
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 8 of 14 CNWRA-14 5.2.1 A height of 15 cm seems non-conservative for large fuel packages that do not consist of thermoplastics that form a pool fire on the floor.
It may be non conservative for some non-thermoplastic packages; however, it is conservative for all thermoplastic packages and will be conservative for some non-thermoplastic packages. We are looking to provide generic guidance and when the types of packages are weighted based on OE against the heights 15 cm is a conservative height. This is demonstrated in the plot below this text (which also addresses the later NUREG 2232 comment about the use of the package half-height). The plot shows the measured flame height from video vs. the flame heights predicted with the FDTs for all packages with a non-zero height. The plot clearly shows that the height of 15 cm results in an overall conservative bias. It also demonstrates that the use of the half height in the weighting process is appropriate as the half height values are all greater than or equal to the measured height.
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 9 of 14 Comment No.
Document Number Section / Paragraph Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Reviewer Acceptance /
Date CNWRA-15 5.2.2 Replace Figure 5-7 and Figure 5-8 shows the yields with Figure 5-7 and Figure 5-8 show the adjusted yields
, assuming the soot yields in Figure 5-7 are normalized to a heat of combustion of 25 MJ/kg Accept comment CNWRA-16 Fig 5-11 Show tg, tp, and td on the figure Accept comment 0
0.5 1
1.5 2
2.5 3
0 1
2 3
FDTS Flame Height (m)
Measured Flame Height (m) z=0 z=half z=0.15
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 10 of 14 Comment No.
Document Number Section / Paragraph Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Reviewer Acceptance /
Date CNWRA-17 Fig 5-12 It is a stretch to consider this a strong correlation but the figure definitely shows a clear trend.
Adjusted wording and added gamma fit to plot to add to argument.
CNWRA-18 5.2.3.2 Replace that high R2 values exists for with that relatively high R2 values exist for Accept comment CNWRA-19 Table 5-3 Replace the symbol for the effective height, z, with ze.
Accept comment CNWRA-20 Page 6-3 A height of 15 cm seems non-conservative for large fuel packages that do not consist of melting thermoplastics (also see comment page 5-7, line 7).
See above response CNWRA-21 Page 7-2 The production rates of soot and CO depend on the soot and CO yields, the heat of combustion of the fuel and the HRR of the fire.
Accept comment CNWRA-22 Table 8-3 and discussion Replace z, with ze.
Accept comment
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 11 of 14 Comment No.
Document Number Section / Paragraph Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Reviewer Acceptance /
Date CNWRA-23 Figures A-1 through A-3 and A-12 through A-17 show that, for the prior testing, the fixed decay exponent of 0.32 generally over-predicts the HRR during the decay phase.
Would this imply that the HRR is generally under-predicted for the tests described in reference [8]? If so, would it not have been better not to include the data from prior testing because our ability to make accurate HRR measurements is much better now than in the days the prior testing was performed? Six tests in Table 2-1 had problems with the calorimeter, rendering the data unreliable (see p. 2-9 and p.
2-10).
The blue dotted line is not the 0.32 exponent but the best fit exponent given in the tables below. This is noted in the test prior to the plots:
Plots of the HRR also show the curve resulting from the fitting fire growth and decay parameters It is recognized earlier in the report that this fitting process doesnt capture the very tail of the fire, however, it is the initial drop off the peak that defines the ZOI and that is generally well captured.
Much better is a relative term. It is not as if we went from 50 % to 10 % error in our ability to do calorimetery. The old tests are not that inaccurate to throw them out altogether.
Tests with calorimeter issues were not used.
CNWRA-24 Editorial comments Generally accepted except as noted below CNWRA-24a 3-2 There appears to be no connection between the numbered list and the preceding paragraphs. The word data is plural, i.e., replace Input data is as with Input data are as,
Input data accounts for with Input data account for etc.
The lead in paragraph says there were three goals.
The list has three main elements each of which describes a goal.
Plural comment accepted.
NUREG 2232: While out of scope and the report has been published, comments are briefly addressed
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 12 of 14 Comment No.
Document Number Section / Paragraph Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Reviewer Acceptance /
Date 2232-1 4.1.1 The 100 kW calorimeter was used for the smaller transient combustibles that were tested. It is stated that a 226.8 kg load cell was used to measure the mass loss of the test specimens. For the smaller items, i.e., those that weighed a few kg or less, it is not clear whether the measurement uncertainty of the load cell was low enough to obtain accurate mass loss measurements. The latter are used to calculate the soot and CO yields and the heat of combustion.
As discussed in the test report, for tests with very low mass loss, yields were not reported. This can be seen in the N/A values in the tables.
2232-2 5.2 This section indicates that the soot mas was determined based on light extinction measurements. The Fire Technology Department at SwRI has the experience that this method generally results in low values for the soot yield.
Reasons are not entirely clear Unable to locate a peer reviewed publication comparing gravimetric to attenuation soot yield measurements for a hood calorimeter. If this conclusion is based on comparison to SFPE Handbook values, I question the validity given that soot yield changes with scale and the SFPE values are all bench scale type measurements. Scale issues aside, the values of soot yield measured do not seem out of line with values reported elsewhere with yields of 1-2 % for cellulosic fuels, yields of 3 to 8 % for TP materials, and 3 to 12 %
for other materials which includes TS materials.
Ignoring the differences in scale the SFPE Handbook shows yields of 1 to 2 % for wood, paper, and carboard and yields of 7 % for PE. The report recognizes that this has more uncertainty than a CO yield due to the assumed extinction coefficient.
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 13 of 14 Comment No.
Document Number Section / Paragraph Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Reviewer Acceptance /
Date 2232-3 5.4 The fire Froude number is a function of the HRR of the fire and the diameter of the fire base. The latter was estimated from the test videos and has a high uncertainty, in particular if the transient combustible is irregular in shape I disagree that the test diameter has high uncertainty at least for the larger fires (>10 kW) used to determine Q* in the PRA report. Fuel packages over 10 kW generally had a uniform area of combustion which in many cases was the footprint of the object making the diameter well known.
2232-4 5.5 It appears that for this comparison a radiative fraction of 30%
was assumed, which may be too low for some transient combustibles that were tested This is the standard assumption made in PRAs and contained within NRC guidance. The FDT vs data heat flux predictions in the report do not suggest that this is a significant bias especially in the range of fires modeled in PRA 2232-5 5.6.3 This section describes the method that was used to obtain the required thermal response data needed for the application of the heat soak method to Kerite FR cable. Draft NUREG-2233 refers to NUREG-2178, Vol. 2, in which the heat soak method for TS and TP cable is discussed. Draft NUREG-2233 should also cite NUREG-2232 to point the reader to a discussion of the adaptation of the heat soak method for Kerite FR cable Added a reference to the test report and NUREG/CR-7102 to discussion of uncertainty and conservatism in the PRA report.
2232-6 5.8 Assuming that the fire elevation is at mid-height of the fuel package may be non-conservative.
See response above
REVIEW / COMMENT DOCUMENTATION Reviewer:
Document #, Rev:
NUREG-2233/ EPRI 3002016054 Discipline/Department:
Various Date:
March 2020
Title:
Transient Fires - PRA Comments shall be:
- COMPLETE AND INCLUDE A REFERENCE TO THE AFFECTED DOCUMENT
- LEGIBLE AND REPRODUCIBLE
- FOCUSED TO A SPECIFIC PROBLEM OR DEFICIENCY Page 14 of 14 Comment No.
Document Number Section / Paragraph Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Reviewer Acceptance /
Date NFPA (Public Comment #1)
NFPA-1 N/A Suggestion to use more NFPA standards in regulation.
The authors of the report recognize that there can be societal benefits when regulatory agencies adopt consensus codes; however, this comment is not germane to the subject matter contained in the report. No changes will be made to report contents based on this comment; however, it will be forwarded to the Office of Nuclear Reactor Regulation for their consideration.