NUREG-2178, Comment Resolution - NUREG-2178 Vol 2 Final

Comment Resolution - NUREG-2178 Vol 2 Final

ML19347B923
Person / Time
Issue date:	08/31/2019
From:	Office of Nuclear Regulatory Research
To:
Dave Stroup 415-1649
References
EPRI 3002016052 NUREG-2178 Vol 2
Download: ML19347B923 (55)

Text

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Introduction & Front Matter CNWRA1 Page 1-2, line 16 Replace from fires with from ignition source fires Revised. Confirmed EPRI report says no cable tray 10/4/19 because the new fire location factors do not apply to fires.

cable tray fires, for example.

NEI-1 1 (pdf pg 30) "HRCNWRRs for electric motors and dry transformers: Text revised as (VLO, 09/16/2019): 10/4/19 Appendix G of NUREG/CR-6850 recommended HRRs for electric motors and dry transformers:

bounding/conservative values for HRRs associated with Appendix G of NUREG/CR-6850 recommended electric motors and dry transformers based on the values HRRs values for electric motors and dry assessed for electrical cabinet fires. " transformers based on the values assessed for electrical cabinet fires.

HRRs from 6850 were not bounding/conservative in all instances, as compared to the values presented in 5, recommend rephrasing this statement.

NEI-2 1 (pdf pg 30) The section for Non-suppression floor value doesn't end The discussion on the NSP floor was moved into 10/4/19 with a period. NUREG-2230.

Add a period to the end of the section on 'Non-suppression floor value'.

NEI-3 1 (pdf pg 30) The discussion on 'Non-suppression floor value' doesn't The discussion on the NSP floor was moved into 10/4/19 indicate that it is only applicable to MCR fires, however the NUREG-2230.

discussion within the report does limit the applicability.

Recommend including some discussion about the non-suppression floor value change being applicable only to the MCR.

NEI-4 1.3 (pdf pg 31; lines 18- The terms electrical enclosures and electrical cabinets, To be revised as: 10/4/19

25) as used in this report, are inclusive of cabinets, panels, and relay racks as those terms are used in NUREG/CR-6850 and other related FPRA documents and standards. In NUREG/CR-6850 the term enclosure is used in an additional two other contexts - [] and the In NUREG/CR-6850 the term enclosure is used in two modeling of enclosure fires (i.e., fires that occur other contextsnamely, the regulatory issue of cables and within a room as opposed to fires that occur in an components that share a common enclosure (e.g., cables open unconfined space). The discussion presented in Chapter 4 of report related to fires Page 1 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date routed in the same cable tray), and the modeling of spreading between adjacent electrical enclosure fires (i.e., fires that occur within a room as enclosures uses the text enclosure fire to opposed to fires that occur in an open unconfined represented electrical enclosure. The reader is space). The reader is cautioned not to confuse these cautioned to consider the context of the unrelated uses of the word enclosure. discussion, specifically to the use of the term This discussion implies that when used, the term "enclosure enclosure.

fires" does not relate to "electrical enclosures" or "electrical cabinets" but instead means a fire occurring in a room.

However, in many places of this report the term "enclosure fire" is directly used to refer to a fire in an electrical cabinet (most often in Section 4).

This section should be revised to clarify that "enclosure fire" is used in 2178, Vol 2 to mean an electrical cabinet fire.

NEI-5 "Section 4 provides Include comma "NUREG/CR-6850, Appendix S" Revised as suggested 10/15/19 additional guidance beyond that in NUREG/CR-6850 Appendix S" Obstructed Radiation (Sections 2 & 3, Appendix B)

NEI-6 Section 2 If available, provide FDS results for obstructed radiant runs The only FDS runs for the items that exist in Table 10/1/19 in excel format. This will limit the use of interpolation the 2-3. The ZOI is the parameter of interest and it is analyst needs to perform and therefore reduce uncertainty. already tabulated. There is nothing that can be interpolated from the FDS runs that isnt in Table 2-3 NEI-7 Section 2 and 3 Provide statement on whether or not obstructed radiation The obstructed radiation factors are applicable for 10/15/19 factors can be used if using the point source method. the solid flame model only. Noted in text, see Section 2.2.4.

NEI-8 2, 3, and Appendix B Obstruction factors greater than 1 will result in ZOIs larger Clarified this in Section 3.1.1 after Figure 3-1, 10/16/19 than those predicted by the FDTs. Clarify whether or not added:

the analyst must use ZOIs larger than those produced by the FDTs. Note that in one case, large, open, TP cabinet with default fuel load, that the Obs_fac is greater than Page 2 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date one. This indicates that for this specific scenario the cabinet ZOI is larger than that for the unadjusted FDTs. If obstructed radiation values are being applied, this result should be preserved and not revert back to the original FDTs value.

NEI-9 2, 3, and Appendix B A simplified approach to obstructed radiant ZOI when using This is a good suggestion. Looks like this is 10/16/19 damage threshold method would be to apply the 98th% universally true and added it as a suggestion.

obstruction factor of a given source to all percentiles of the fire since the 98th% obstruction factor will always be Revised text is in Section 3.3.1 after Figure 3-1:

bounding. Suggest including this method as a suggested Section 3.1.1 after Fig 3-1 added:

approach. It is noted that the 98th percentile Obs_Fac is bounding for lower percentiles. Therefore one can get a conservative estimate on the ZOI for percentiles other than the 98th percentile values in Table 3-1 and Table 3-2, by computing the ZOI using the unadjusted solid flame model and applying the 98th percentile ZOI. Conversely, one could also use the values shown in Table 2-6.

NEI-10 2, 3, and Appendix B A simplified approach to obstructed radiant ZOI when using See above (duplicate comment) 10/15/19 damage threshold method would be to apply the 98th%

obstruction factor of a given source to all percentiles of the fire since the 98th% obstruction factor will always be bounding. Suggest including this method as a potential approach.

NEI-11 2, 3, and Appendix B Chapter 2 establishes that the adjusted FDT is the Industry members of the working group requested 10/16/19 suggested radiant ZOI model for unobstructed radiation. the current format. The rationale was that many However, in the main body sections pertaining to obstructed plants already have tables for the current FDTs and radiation, the values presented use the unadjusted FDT and having the report body reflect that would be the adjusted FDT values are maintained in App. B rather preferable.

than the main report. Furthermore, the information provided in App. B is limited compared to that in the main report.

Suggest that for the damage threshold approach It is only the Obs_fac that is dependent on documented in the main report, the adjusted FDT be used adjusted vs. unadjusted. The severity factors dont and App. B should document the unadjusted FDT. Both change. This can be noted in the App. Added the Page 3 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date methods should include full analysis (e.g., obstruction equivalent tables (Table B-3) for Groups 1-3 for factors, max severity factors, Groups 1-3 approach, etc.) Obs_fac CNWRA2 Page 2-1, lines 13 Replace discussed with critically examined Edit text to change discuss to examine 10/3/2019 because the underlying assumptions have been discussed during the NFPA 805 LAR reviews but have not been examined in any detail.

CNWRA3 Page 2-1, line 35 Replace currently used in FPRA applications. with Revised text to: 10/3/19 currently used in FPRA applications, i.e., for fires that are an order of magnitude smaller in physical size. The next two sections investigate the FDT s correlations as they are currently used in FPRA applications. The fires postulated in a FPRA are typically an order of magnitude smaller in physical size.

CNWRA4 Page 2-2, line 12 What is near zero? Replace The view factor of the fire The flame size is small is not the underlying 10/3/19 from the target is near zero with The flame size is small mathematical assumption. The underlying relative to the distance between the target and the fire mathematical assumption is that mathematically the fire is a point or equivalently the view factor has approached zero. Edited to say 2 The fire can be considered a point with zero surface area when computing the radiation view factor. This requires the target be located at a sufficiently large distance from the fire.

CNWRA5 Page 2-2, line 28 It is common to assume a default value of 0.3 for the It should be noted that values from App A of the 10/3/19 radiative fraction. However, this may be too low for cable SFPE handbook are primarily from small scale fire fires and other types of fires that need to be considered in tests that may not be larger enough to be fully an FPRA. For example, the radiative fraction for cable fires turbulent as we would expect for the cabinet fires can be estimated from the ratio of the radiative to the that pose a significant hazard to radiant targets.

chemical heat of combustion reported in Table A.39 of the The soot yields and radiant fraction may not scale.

SFPE handbook (5th Edition). The resulting range for the Some of the high values reported seem radiative fraction of TP cables (PE/PVC and inconsistent with current knowledge of extinction PVC/nylon/PVC-nylon) is 0.37-0.63 with a mean of 0.50 and limits. A radiant fraction of 60 % would imply a median of 0.49. For TS cables (EPR/Hypalon, XLPE/XLPE, flame temperature of ~1000 K. Methane at its LFL and XLPE/neoprene) the range is 0.34-0.62 with a mean of needs ~1200 K to sustain a flame.

0.47 and median of 0.50. Consequently, a single range of Page 4 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date 0.35-0.63 for the two cable types seems reasonable. This is The purpose of Table 2-1 is to demonstrate that, well above the commonly used value of 0.3. As a result, the as typically used in FPRA, targets at the ZOI are target distances in Table 2-1 may be too low (which is non- not far enough away for the assumptions of the conservative). point source model. Typical practice is to use the default value of the radiant fraction built-in to the Note: Using a low value for the radiative (or radiant) fraction FDTS.

may have an effect on model results and conclusions presented in other sections of the NUREG. A more detailed examination would be needed to quantify this effect and to In the FDS modeling, in most configurations the determine whether additional calculations with a higher hazard is not direct flame radiation but rather the radiative fraction need to be performed. wall surface temperature of the cabinet. This is less sensitive to the radiant fraction.

CNWRA-5 Follow-up: We recommend that the Office of Regulatory Research consider conducting a study to obtain more accurate values for the radiative fraction and soot yield of various types of electrical cables.

The Office of Nuclear Regulatory Research will consider conducting a study to obtain more accurate values for the radiative fraction and soot yield for various types of electrical cable in the future as available resources permit.

CNWRA6 Page 2-3, Table 2-1 Apparently, Heskestads flame length correlation was used Added 10/3/19 to calculate Lf. That should be stated in the text or a footnote to the table.

NEI-68 Table 2-1 The number of significant figures shown varies within a Adjust sig figs to same number 10/3/19 column. For example, the flame height of the small cabinet Page 35 reports the measurement in meters to the hundredths place and the 378.5 liter oil spill only reports to the ones place.

Page 5 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Make number of significant figures shown within a given column consistent.

NEI-69 Table 2-1 The Distance for Heat Flux values presented do not take Added note in text that distances are total. 10/3/19 into account the fire diameter, D. Based on the equation (Page 35) presented in Figure 2-1 and FDT 05.1, the values calculated should have D/2 subtracted. Update values OR add assumption that the area is assumed 0.

If fire diameter is included, update Section 2.2.1 pertaining to Distance Ratio to discuss highest ratio of 1.93 following the update to Figure 2-1.

CNWRA7 Page 2-3, lines 10-11 It is suggested to revise the last sentence. Wind effects are Edited 10/3/19 (usually) not accounted for in FPRA applications. However, in addition to target orientation the view factor calculation also depends on the physical size of the fire and target distance to the fire.

CNWRA8 Page 2-4, Figure 2-3 S should be drawn as a cylinder. Updated figure 10/3/19 CNWRA9 Page 2-4, lines 19-21 When determining the radiant fraction (e.g., from Revised 10/3/19 experiments), the radiative heat transfer to the fuel (i.e.,

through the bottom surface of the cylinder) is not included.

It is therefore suggested to replace the sentence at the bottom of page 2-4 with the following: The portion of the surface integral in Equation 2-1 over the bottom of the cylinder can be ignored because in determining the radiant fraction (e.g., from experiments), the radiative heat losses to the fuel (i.e., through the bottom surface of the cylinder) are not included. Furthermore, assuming all thermal radiation emitted by the flame passes through the side of the cylinder, the result of integration is shown in Equation 2-2. This assumption is conservative because it results in a higher value for the emissive power, E, and therefore a larger ZOI.

NEI-12 2.2.2 (pg 38) Equation 2-2 Equation corrected. Figure was correct. 10/3/19 The surface, S, is dependent on LF and D, not height of target above fire base, H. Equation 2-4 correctly shows this.

Page 6 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Update equation 2-2 to replace H with LF. Ensure the graph in Figure 2-4 is using the correct equation.

CNWRA10 The part of the paragraph starting with The emissive power Removed and replaced with Using Equation 2-4.. 10/3/19 and ending with periphery of the fire is confusing Page 2-6, lines 13-16 and redundant. It is suggested that these sentences be deleted.

CNWRA11 Replace a 30% radiant fraction with an assumed Edited. 10/3/19 Page 2-6, line 21 30% radiant fraction Also, see comment on Page 2-2, line 28 CNWRA12 For consistency it is suggested to use the same term when Edited 10/3/19 referring to the heat flux threshold for damage. For example, in line 30, replace a higher heat flux with a higher critical target heat flux . In line 31, replace Page 2-6, lines 30-36 desired heat flux with actual critical target heat flux

. In line 33, replace of the desired heat flux. with below the actual critical target heat flux. In Equation 2-6 use q _(cr,eff)^" and q _cr^" to refer to the adjusted and actual critical target heat flux, respectively.

NEI-65 The ratio of these two is 2.2. The 2.2 is a result of doing the calculation without 10/3/19 rounding the intermediate steps. However, it is Section 2.2.3.1 (pg 38) The ratio of the two is 2.1 (57.5/27.3). The 2.2 value may correct that using the rounded intermediate values have mistakenly been taken from the area of the cylinder. would give 2.1. For clarity I agree that using 2.1 Update paragraph to use 2.1. would be better.

CNWRA13 Same as previous comment: Fixed 10/3/19 For consistency it is suggested to use the same term when referring to the heat flux threshold for damage. For example, in line 30, replace a higher heat flux with a higher critical target heat flux . In line 31, replace Page 2-7, line 2 desired heat flux with actual critical target heat flux

. In line 33, replace of the desired heat flux. with below the actual critical target heat flux. In Equation 2-6 use q _(cr,eff)^" and q _cr^" to refer to the adjusted and actual critical target heat flux, respectively.

Page 7 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date NEI-66 The change to emissive power does not change 10/3/19 This results in horizontal ZOIs of 1.65 m (5.4 ft) and 0.93 m how a plant would determine the ZOI. We are not (3.1 ft) from the edge of the fire respectively for the TP (6 suggesting any change to any existing method kW/m2) and TS threshold damage fluxes (11 kW/m2). other than if the method is using an emissive power based on the Shokri-Beyler correlation then It is not clear how the ZOI is to be calculated once an either the emissive power should instead be Section 2.3.1 (pg 44) adjusted emissivity is calculated. Clarify that when using computed per Equation 2-5 or the target heat flux method 2 of Section 2.2.3.1 (i.e., heat flux ratio), the analyst be adjusted per Equation 2-6 (whatever is easier should use the Solid Flame 2 tab of FDT 05.1 and assume based on the method being used by the plant).

the target is located at a height of 1/2 the flame height (i.e., How a plant ultimately uses computed heat flux bounding case). (threshold value, heatsoak, THIEF etc.) is not relevant to this section.

CNWRA14 Page 2-13, lines 6-9 Equation 2-7 is incorrect. The terms inside both Yes the amb and w should be reversed. Also good 10/3/19 parentheses on the right-hand side of the equation should point that the effective flame temperature could be be reversed. For example, (Tamb - Tw) should be (Tw - replaced by the emissive power. These changes Tamb). In addition, there is no need to introduce an were made.

effective blackbody radiation temperature by writing the equation as follows:

in (E Tw4 ) + hin Tg Tw = (

4 out (Tw4 Tamb ) + hout (Tw 2 Tamb ) -

7

)

CNWRA15 Page 2-13, line 11 35 W/m2K seems reasonable, but the cited reference Changed to a Quintiere paper and made second 10/15/19 (Eurocode 1-2) deals with fires in structures and does not edit.

seem to be applicable here. A better reference is needed.

Page 2-13, lines 13-15 In line 13, replace Radiative flux with Maximum radiative flux because Equation 2-8 is for a target that is facing the center (mid-width and mid-height) of the cabinet panel. Also, use out instead of in the equation.

Page 8 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date NEI-13 2.3.2 (pg 46) ...size for the cabinet size and door status based on Edited 10/3/19 NUREG-2178, i.e., 325 and 1,000 MW for medium and large cabinets, respectively.

Change units from MW to kW.

CNWRA16 Page 2-14, lines 9-11 From the description it appears that the fire consisted of a Added 10/3/19 gas burner, but that is not stated anywhere.

CNWRA17 Page 2-14, lines 19-24 Explain what K is in Equation 2-9 and provide the value that Fixed 10/3/19 was used for K. The units for the convection coefficient in line 23 are W/m2K, not W/mK.

CNWRA18 Page 2-18, lines 16-20 Guidance for the treatment of different types of Kerite cable The guidance is for threshold values based on 10/3/19 in terms of damage and ignition thresholds is provided in the temperature. This work is primarily focused on time NRC Fire Protection Significance Determination Process dependent exposures using the heat soak method (see reference [54], pp. 64-65). This guidance is based on for which no kerite equivalent data to the 6850 experience from NFPA 805 LAR reviews. tables of flux vs damage time exist. Given that the recommendation to use TP as a surrogate is applicable.

CNWRA18 Follow-up: We propose replacing the last two sentences at the end of the paragraph below the two bullet items at the top of page 2-18 (Starts with Since the guidance for Kerite-FR cables in ) with the following: Based on the test results in NUREG/CR-7102 and guidance in FAQ 08-0053 Revision 1, TP damage criteria can be assumed for Kerite-FR cable and TS damage criteria can be assumed for Kerite FR-II, FR-III, and HT cable. Furthermore, TS ignition criteria can be assumed for all Kerite cable types because all varieties are IEEE 383 qualified.

Page 9 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Response to follow-up: Added except for the last sentence. This report only deals with the damage ZOI and not ignition.

NEI-14 2.3.3 (pg 50) Vented / unvented face - A vented cabinet face is a face In the bulleted text after Figure 3-1 added: 10/16/19 where there are openings to support substantial air flow in or out of the cabinet. These would include louvers for A cabinet face with a non-robustly secured door passive or mechanical ventilation, areas of wire mesh, large (per the definition in Section 6.5.6 Step 6 Bin 15 of areas with no panel present, or a face with an access panel NUREG/CR-6850 [1]) would be considered a

/ door that is not robustly secured. vented face. An unvented face is any face that is not vented. Note that an unvented face can contain The final criteria implies that any panel that does not meet small openings for cable feed throughs.

the definition of robustly secured is to be considered vented.

Define robustly secured (e.g., guidance from FAQ 08-0042).

NEI-15 2.3.3 (pg 50) A vented cabinet face is a face where there are openings to In the bulleted text after Figure 3-1 added: 10/16/19 support substantial air flow in or out of the cabinet. These would include louvers for passive or mechanical ventilation, A cabinet face with a non-robustly secured door areas of wire mesh, large areas with no panel present, or a (per the definition in Section 6.5.6 Step 6 Bin 15 of face with an access panel / door that is not robustly NUREG/CR-6850 [1]) would be considered a secured. An unvented face is a face with no significant vented face. An unvented face is any face that is openings to support substantial air flow in or out of the not vented. Note that an unvented face can contain cabinet. small openings for cable feed throughs.

The final criteria for a vented face implies that a panel that is well sealed, but not necessarily robustly secured (e.g., has a simple twist-handle style top-and-bottom- latches), is to be considered vented. However, the next sentence suggests that if the face is unvented (i.e., well sealed), the panel can be considered unvented. Provide clarification on the treatment of well sealed panels that do not meet the definition of well secured.

NEI-16 2.3.3.2 (pg 54) For the south and west faces the FDS predicted ZOIs are Fixed 10/3/19 less than the adjusted FDTs ZOIs for 85% to 92% of the results. 3% to 8% of the FDS ZOIs are greater than 110%

of the adjusted FDTs ZOIs.

Page 10 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date These statistics also apply to the east wall. Update wording to include the east wall.

CNWRA19 Page 2-28, lines 10-15 In lines 14-15 it is stated that the medium grid was used for No. The increase in flux to the medium is mostly 10/3/19 the simulations. However, in lines 10-11 it is states that compensated for by the bias shown in the there was a 20-30% increase in the predicted heat flux simulations of the NIST tests. Combined with the going from the medium to the high resolution. Was a 1.2- conservative bias shown in the heat soak method, 1.3 correction factor applied to the heat flux predictions used there is no need to add on additional in generating the results that are presented in Tables 2-3 conservativism. Added a note on the heat soak through 2-6? conservatism.

CNWRA-19 Follow-up: We suggest expanding the discussion in section 2.3.3.3 to better show that the bias in the FDS heat flux calculations combined with the bias in the heat soak method indeed compensates for the error in the heat flux predictions due to the use of the medium resolution grid. More specifically, an estimate of the bias in the heat flux method should be provided and reference should be made to section A.5.

Response to follow-up: A grid sensitivity study was performed using a 0.3 m3 (12 ft3), medium cabinet with a 75th percentile, steady-state fire in the central cylinder configuration. The baseline grid size (medium resolution [MR]) was decreased by 50 % (low resolution [LR]) and increased by 50 %

(high resolution [HR]). Runtimes were approximately 0.15, 2.4, and 27 CPU days for the three grid resolutions. There is a significant increase in the predicted flux from the low to the medium resolution and lesser increase (20 to 30

%) from the medium to the high resolution. Results of modeling the NIST tests, which use the medium grid resolution, show a 10 % positive bias. This Page 11 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date indicates that the use of the medium resolution would result in a positive bias and high resolution a larger positive bias. Additionally, it is noted that the heat soak method discussed in Appendix A also shows a positive bias. Of the nine tests shown in Figure A-6, five show a faster time to damage, one shows a slower time to damage, and three show an equivalent time to damage. Note that bias in time to damage doesnt directly correlate to bias in heat flux due to non-linear response as a function of heat flux. compensation will occur based on the heat soak method discussed in Appendix A. Given that the NIST validation exercise shows a positive bias for the medium grid, that additional bias will result from the heat soak method, and the computational resource requirements for the various grid resolutions, the medium grid was selected for use in simulations. Figure 2-26 shows the grid study results of predicted heat flux as a function of distance from the cabinet for each cabinet face at three vertical locations: near the top of the cabinet, near the bottom of the cabinet, and at the mid-height of the cabinet.

CNWRA20 Page 2-31, lines 13-17 Was the heat soak method described in Appendix used to No. This was simply done as a threshold exposure 10/3/19 determine the plume ZOI? as noted in the text before Figure 2-38.

NEI-70 Table 2-5 The Adjusted FDT ZOI values given for the 1000kW fire do The example in 2.3.1 is a different cabinet (63 ft3 10/3/19 not match with those provided in Section 2.3.1. Make vs 50 ft3). The values should be different.

(Page 67) values consistent.

CNWRA21 Section 3 (General) For large cabinets with an open top it may be possible for No this analysis does not account for a cabinet 10/3/19 combustion to take place above the cabinet. Is the radiation with an open top. This is a good observation and from such a flame accounted for in the analysis? text was added to the beginning of Section 3 to note that if a cabinet has an open top that the normal approach for an open fire should be applied.

Page 12 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date 2 (pg 33) a solid flame model [3], or the use of field model such as Fixed (actually appears at end of first paragraph in 10/3/19 NEI-17 the Fire Dynamics Simulator (FDS) [4, 5]. Section 2).

Update wording:

a solid flame model [3], or the use a of field model such as the Fire Dynamics Simulator (FDS) [4, 5].

NEI-18 Figure 3-1 Figure 3-1 implies that there is no benefit on the vented Figure edited 10/3/19 cabinet face, but it should be implying less benefit. The green and red lines should be separated on the vented face.

NEI-19 3.1.1 Provide definitions of Threshold Approach ZOI Factor or Added after Figure 3-1 10/3/19 Damage Integral Approach ZOI Factor prior to Table 3-1 or point to Section 2.3.3.4.

BJ-1 Page 3-3, Table 3-1 Typo in last columns: unnecessary r in Damage Integral Fixed 10/3/19 and Table 3-2 BJ-2 Page 3-3 ZOI factor of very low is greater than low for Large- Obs_fac is not the ZOI. It is the multiplier of the 10/3/19 Closed-TP (Damage Integral Approach) and Medium- solid flame ZOI. There are multiple non-linear Table 3-2 Closed-TS (Both Approaches). factors that impact the amount of reduction ZOI that is seen. Even though the obs_fac may be If they are correct, suggest explaining why these very low larger for the cases you cite, the end ZOI that cases have greater factors than low cases. results will still be larger for low than for very low.

NEI-20 3.1.1 The values within Table 3-2 for TP cables result in less See BJ-2. 10/3/19 Table 3-2 severe consequences than TS cables for large, closed (0.76 vs 0.78) and medium, open (0.9 vs 0.98) cabinets.

Thermoset is worse than TP for Medium Open and large closed for default loading. Confirm this is correct and recommend including a note below the table to verify.

CNWRA22 Page 3-4, 3.1.2 For the two examples, add references to the tables (and Example discussion expanded 10/3/19 Example documents if not in the present document) where the 98th percentile HRR, unadjusted radiation ZOI, and Obs_Fac can be found. For example, the second sentence would be Page 13 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date modified as follows: This would have a heat release rate (HRR) of 45 kW at 98th percentile (from Table 4-2 in NUREG-2178, Vol.1).

NEI-21 3.1.2 (pg 74) Provide all inputs necessary to determine the ZOI of the See CNWRA22 example sources.

NEI-22 3.1.2 (pg 74) Identify if the Threshold Approach ZOI Factor or Damage See CNWRA22 Integral Approach ZOI Factor is being used in the example.

CNWRA23 Page 3-7, 3.2.2 This example is difficult to follow. From the first sentence, Edited example to correct values and provide more 10/3/19 Example we can determine that we should be able to find the ZOI in detail.

Table 2-6 because that is the table to be used for large closed cabinets. However, the description of the cabinet in the first sentence does not provide enough information (fuel content and type of cables not specified) to determine the ZOI. The second sentence indicates that both the unvented face damage threshold and integral approach ZOI are equal to 0.59 m. According to Table 2-6, a 285 kW fire is the only fire in a closed large electrical cabinet that results in a ZOI of 0.59 m (both approaches) for a TP target. At this stage we know that the fuel loading is the default loading but we still dont know whether the cabinet contents are TP or TS cable. From the fourth sentence and Table 3-3 we can then determine that the cable type is TS (because the maximum severity factor is 0.10). Finally, Table 2-6 indicates that a TP target at 0.25 m from an unvented face of the cabinet with TS contents would still be damaged at an HRR of 202 kW (93rd percentile HRR) but not at 155 kW (90th percentile HRR). Consequently, the severity factor for the cable is between 0.07 and 0.1 and slightly higher than 0.07.

It would be helpful to describe this example in more detail.

In addition, it is not clear at the start of the example what the purpose is of this example. It becomes clear later on that the example provides a comparison between the maximum Page 14 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date severity factor in Table 3-3 or 3-4 and the more accurate value determined from the ZOI tables in Section 2.

NEI-23 3.2.2 (3-7) "Consider a TP electrical cable that is located 0.25 m (0.8 ft) See previous comment. 10/3/19 from the unvented face of large, closed cabinet. For this cabinet Table 3-3, has a maximum severity factor of 0.10.

The actual severity factor for the cable could not be larger than 0.1; however, it could be less. If the detailed tables from Section 2 were used, the severity factor for this cable would be 0.07.

This example states the maximum severity factor is 0.10 or 0.07. This is for low fuel loading large cabinets, which is not stated in the example. For default fuel loading cabinets, the maximum is 0.25/0.20.

Specify the appropriate fuel loading in the example.

BJ-3 Table 3-6 While the severity factors of thermoset cables are equal to Just because the 98th percentile HRR is greater 10/3/19 or less than thermoplastic cables as expected, there are does not mean that the gamma distribution is Table 3-3 and Table 3-4 inconsistencies between Open and Closed cabinets as well greater for all percentiles. In the cases referred to as between Large and Medium cabinets. the medium closed is actually more severe at lower percentiles which results in a greater severity For examples, for vented cabinet (Table 3-4), severity factor factor.

of medium-closed-TP is greater than SF of medium-open-TP and Large-closed-TP.

Note that Table 4-2 of NUREG-2178 Vol. 1 shows that HRR of medium size cabinet is equal to or less than HRR of large cabinet and HRR of Open cabinet is greater than Closed cabinets.

NEI-24 Tables 3-3 and Tables Regarding Tables 3-3 and 3-4, typically severity factors less Replaced with SCRN 10/3/19 3-4 than 0.02 are screened. These tables have values of less than 0.02 (e.g. 0.01). Recommend using a screen abbreviation when there is no external impact (e.g.

Page 15 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date SCRN). This could also be used in Tables 3-1 and 3-2 for clarity.

NEI-71 Table 3-3 and Table 3-4 It is not clear which FDT (i.e., adjusted or unadjusted) was These tables are based solely on the FDS 10/3/19 (pg 76) and Table 3-9 used to populate the numbers in Table 3-3, Table 3-4 and simulations. A note is added before the tables.

(pg 82) Table 3-9. Provide documentation of which FDT was used.

If unadjusted FDT is used, provide Maximum Severity Factors for adjusted FDT in Appendix B.

CNWRA24 Page 3-13, line 28 Replace for the medium cabinet. with for the small Fixed 10/3/19 cabinets.

NEI-58 Appendix B Do these values apply to Groups 1, 2, and 3 as well as Edited table headings and added in table B-3 for 10/3/19 Groups 4a and 4b? There does not appear to be enough groups 1, 2, and 3. See comment NEI-11 information to determine which groups these apply to.

Please clarify which groups these apply to and if not Groups 1, 2, and 3 then include updated tables for those groups.

CNWRA95 Page B-1, Table B-1 Using the threshold approach for open TP and TS cabinets, Obs_fac is not the ZOI. With this table, the ZOI for 10/3/19 the Obs_Fac for default fuel loading is equal to or smaller a cabinet is given by the ZOI computed using the than the value for low fuel loading. For all other cases in solid flame model times the appropriate Obs_fac.

Table B-1 and all cases in the corresponding Table 3-1 for The Solid flame ZOIs for the default fuel loads are the unadjusted FTDs, the Obs_Fac is larger for default fuel larger than the solid flame ZOIs for the low and loading compared to low loading. Is this a typo or is there a very low fuel loads. The actual ZOIs after plausible explanation for the reversal? multiplying by obs_fac are still larger for default than low.

CNWRA96 Page B-2, Table B-2 Same as previous comment but for vented cabinets faces of See above comment. 10/3/19 large open cabinets with TP cable contents. The corresponding Obs_Fac values for the unadjusted FTDs are provided in Table 3-2.

Section 4 (Cabinet to cabinet, Appendix D)

NEI-25 4.1.4 (pg 88) "No modifications are intended to the existing guidance for Yes, the guidance provided in this section is limited 10/4/19 determination of functional damage to equipment due to fire to determining and modeling fire spread to affecting cables and components inside the electrical adjacent cabinets.

enclosures."

Page 16 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Recommend addressing the following portion of cabinet to Text revised as:

cabinet damage approach discussed in Section S.2 of NUREG/CR-6850: "Assume no damage in the second Time to ignition should be considered in adjacent cabinet occurs until after the fire propagates to the damage estimates. But, no changes in the rules adjacent cabinet. Assume damage can occur earlier if there governing the functional damage to equipment are large openings in a wall and plenum areas in which a due to fire affecting cables and components are hot gas layer is likely to form." suggested.

If damage is not postulated until after ignition, and the rules for ignition are now changing, confirm that the new rules for ignition do not invalidate the previous rule for damage based on ignition.

NEI-26 4.1.4 (pg 88) "3) The guideline and methods described here are not HEAF is the subject of ongoing research. This 10/4/19 intended to apply to high energy arc fault (HEAF) fire guidance is only applicable to thermal fires.

scenarios." Text revised as:

Can this be elaborated on? As in, what guidance should be 3) The guideline and methods described here is followed for HEAF scenarios or is this information expected limited to fire spread between adjacent to be included in future testing? cabinets due to thermal fires and are not intended to apply to high energy arc fault (HEAF) fire scenarios.

NEI-27 4.2.2.1, 4.2.3, & 9.2 4.2.2.1: "Double Wall" Revised as suggested 10/4/19 (Various pages) 9.2: "Cabinet to cabinet fire propagation can be screened for the following configurations: Reviewed for [..] air gap and deleted as necessary

• Double wall air gap" Section 9.2 configuration classification should be changed from "Double wall air gap" to "Double Wall" for clarity and consistency with Sections 4.2.2.1 & 4.2.3.

NEI-28 4.2.2.2 (pdf pg 90) "As a result, for open-top enclosures fire spread along Revised: 10/4/19 horizontal cable runs is considered unlikely."

Page 17 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date This statement seems to be referring to fire spread along As a result, for open-top enclosures fire spread horizontal cable runs from one cabinet to another, however along horizontal cable runs within the cabinet is that is not explicitly clear and could be misconstrued to considered unlikely.

mean open top enclosures wouldn't ignite horizontal runs of cable that are outside the cabinet. Recommend adding some clarifying words to this statement.

CNWRA25 Page 4-5, lines 19-20 This statement is generally true, except for some types of Revised as: 10/4/19 Kerite cable (see reference [54], pp. 64-65). The radiation zone of influence (ZOI) results in Section 2 show no external ZOI for very low fuel load enclosures. Bulk ignition of cable trays will not occur given the lack of an external ZOI.

NEI-29 4.2.3 (pdf pg 94) Recommend revising the terminology of the following Revised to match NUREG-2178 Vol.1 10/4/19 enclosure classifications, as listed in 4.2.2.1 and/or 4.2.3: Classification Groups:

• Small Enclosure

• Switchgear Group 4c: Small enclosure, For consistency and additional clarity revise the enclosure classifications to match between Sections 4.2.2.1, 4.2.3, Group 1: Switchgear and Load Center and 9.2. Group 2: Motor Control Center

• Group 4c: Small Enclosure

• Switchgear & Load Centers Otherwise, "small enclosure" could be mistakenly be thought to be engineering judgment.

CNWRA26 Page 4-10, lines 22-32 It is stated that there were zero occurrences of cabinet to Section revised using combination counts of 10/4/19 cabinet fire spread in 130 cabinet fire events. Based on this events in 2169 and 2230. NUREG-2230 adds observation, the likelihood of multi-enclosure fire experience additional electrical cabinet data up through the is estimated at 0.5/131 = 0.0038 because the 131 events year 2014. The total event count was revised to include fires in electrical enclosures for which fire spread to 151, and then reduced to 107 with the removal of adjacent enclosures would not be postulated based on the MCC/SWGR events. Resulting section values guidance in Section 4.2.3. Consequently, the 0.0038 become 0.5/108=0.0046, 0.0046/0.23=0.02.

estimate is low and to obtain a more accurate estimate the denominator in the equation (131) should be reduced by subtracting the events that involved electrical enclosures and configurations for which fire spread would not have Page 18 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date been postulated. It is probably not possible to identify all these events from the information in the FEDB but at least the denominator could be reduced by subtracting the number of switchgear and MCC fires.

CNWRA27 Page 4-12, lines 11-16 Postulating fire spread to either side with a conditional Text revised to provide guidance on when the 10/4/19 probability of 0.01 may be more practical but could equal split should be used and when the total significantly underestimate the risk. probability (0.02) should be assigned to a single cabinet.

Revised as below in RED Fuel loading in one of the exposed electrical enclosure is low (low fuel/steel partition rule, Section 4.2.2.5).

Victor, is the second bullet missing a word (see my suggestion in the underline)

Fuel materials (e.g., cables or components) in one of the exposed enclosure that are mounted (more likely to spread) or are not mounted (less likely to spread) in contact with the partition wall (low fuel / steel partition rule, Section 4.2.2.5)

CNWRA28 Page 4-13, lines 1-10 Most figures in Section 9.2.3 in NUREG/CR-7010, Vol. 1 do The discussion has been revised to discuss the 10/4/19 not show that the fire intensity (HRR) reaches a peak value behavior captured in the model that as fire spreads that is roughly steady over time. along a single cable tray, the overall fire intensity tends to reach a peak and be roughly stead over time.

CNWRA29 Page 4-13, Section NUREG-2230 introduces the concept of interruptible cabinet Revised as: 10/4/19 4.4.2 fires. How would these fires affect the guidance provided in

• Only the growing fraction of fires should be Section 4.4.2? modeled to spread to an adjacent cabinet when propagation is postulated. The interruptible classification of fires, as described in NUREG-2230, was developed to capture the operational Page 19 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date experience that a significant fraction of fires are detected and suppressed with a minimal effort prior to damage to any external target occurs.

Additionally, the probability that the interruptible fraction of fires is not suppressed prior to propagating is very low (on the order of 7E-07) concurrent with the start of growth at 10 minutes.

NEI-30 4.4.2 (pdf pg 99) "If additional resolution of HRR is required for the exposed This guidance has been removed. 10/4/19 cabinet, the exposed cabinet have HRRs sample from the upper 50 % of the exposed cabinet distribution. For example, if the exposed cabinet is a large, open, Figure 4-1 caption revised as:

thermoplastic (TP) enclosure it could be represented by a Total HRR for Fire Spread Between the Growing 1000 kW fire with a severity factor of 1 or as a 1000 kW fire Fraction of an Exposing Group 4a, Open, with a severity factor of 0.25 and 392 kW fire with a severity Thermoplastic, Default Fuel Load Enclosure and factor of 0.75 (where 392 represents the upper 75 % of the an Adjacent Group 4a, Closed Thermoplastic, upper half of the distribution or the 87.5th percentile fire)." Default Fuel Load Enclosure As written, this is difficult to understand. It is not clear what is meant by 'if additional resolution of HRR is required.'

Recommend rewording to better clarify that the user has the flexibility to apply any HRR to the exposed enclosure HRR as long as the HRR is greater than the 50th percentile HRR using the gamma distributions assigned in 2178.

Also, consider including an example of this into Appendix D for additional clarity on the application.

NEI-31 Section 4.4.2 For Section 4.4.2, NUREG-2230 growth profiles should be Revised as: 10/4/19 mentioned as an option. The exposed cabinet growth The exposing enclosure is postulated to follow a should begin 10 minutes after growth starts in the exposing HRR profile consistent with the growth profiles cabinet. as described in NUREG-2230.

[]

Page 20 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date The exposed enclosure should be modeled following the Growing Fire HRR profile described in NUREG-2230 [].

We will assume that if propagation is postulated, since its coming from the result of a 98% fire that hasnt been suppressed for up to 10 minutes after growth has started, the ignition source for the exposed cabinet would better represent that of a growing fire.

CNWRA30 Page 4-13, Section NUREG-2230 introduces the concept of interruptible cabinet Revised to: 10/4/19 4.4.2 fires. How would these fires affect the guidance provided in

• Fire spread to an adjacent electrical enclosure, Section 4.4.2? where possible based on the go/no-go criteria, should be assumed to occur 10 minutes after the start of the growth period of the exposing enclosure for the growing HRR profile as described in NUREG-2230. Therefore, the exposed enclosure will begin its growth stage 10 minutes after ignition of the exposing enclosure.

CNWRA31 Page 4-15, line 41 Add a sentence to clarify that for an interior cabinet in a Text revised as: 10/4/19 bank, fire spread can be postulated to the adjacent cabinet 3. Fire spread will be limited to one adjacent on either side but not to both adjacent cabinets in one enclosure. For an interior cabinet within a bank, scenario. HRR calculations should be postulated to a single cabinet on either side of the exposing cabinet, but not to both adjacent cabinets.

CNWRA32 Page 4-16, Figure 4-1 At the end of the steady burning period the HRR is Figure revised 10/4/19 supposed to abruptly transition to a linear decay. In this figure the HRR gradually transitions from the steady burning period to the linear decay (it looks like the smoothed line was turned on in Excel). Please correct.

Page 21 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date NEI-32 4.5 (pg 4-16) Assume the peak fire intensity for the exposing enclosure This text was added to remove confusion on which 10/4/19 corresponds to the 98th percentile of the peak HRR HRR percentage should be associated with the distribution applicable to the exposing enclosure (i.e., based exposing cabinet.

on size, function, and/or fuel loading conditions). Edited the wording to state:

As written, this sentence states that the peak fire HRR from the exposing enclosure corresponds to the HRR of the The exposing fire is at the 98th percentile of the exposing enclosure, which is redundant. Consider peak HRR distribution applicable to the exposing rephrasing to remove the redundancy. enclosure (i.e., based on size, function, and/or fuel loading conditions).

NEI-59 6 "Ignition in the Exposing enclosure (Enclosure 5) will be Revised example to 200 kW HRR to match 10/4/19 assumed to take places at time = 0. The peak HRR of 325 description of the enclosure kW will be reached after 12 minutes following a t2 growth profile per NUREG/CR-6850."

Enclosures 4, 5, and 6 are discussed in this section, each of which are said to be 2178, Group 4b enclosures with default D.2.2 (pdf pg 242) loading and based on Figure D-1, these enclosures would represent closed door configurations. However, the HRR that is identified is 325kW, which is the HRR for Group 4b, default loading with open doors.

Either fix this by updating the HRR used in the example to match closed door configuration, or update the cabinet to specify that the cabinet has open doors.

CNWRA97 Page D-2, line 7 The vent approximately 0.3 m above the bottom of the Revised. 10/31/19 enclosure in Enclosure A is not shown in Figure D-1.

CNWRA98 Pages D-6 through D-9, At the end of the steady burning period the HRR is Figures revised to show linear decay. 10/4/19 Figures D-2 through D- supposed to abruptly transition to a linear decay. In the 6 figures, the HRR gradually transitions from the steady burning period to the linear decay (it looks like the smoothed line option was turned on in Excel). Please correct.

CNWRA99 Page D-14, line 4 The time to failure would be approximately 19 minutes if the Text revised from damage to failure. 10/4/19 HRR evolution in Figure D-2 is applicable to this example.

Page 22 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Section 5 (Motor & Transformer HRRs)

CNWRA33 5 (General) This section describes several Monte Carlo analyses in Text revised as: 10/4/19 which a set of input parameters are varied in a random manner within a specified range. For most parameters, no Based on the photographic reviews, the fraction basis is provided for the range that was chosen. For is postulated to range between 40-80% of the example, on page 5-8, line 21 it is stated that the fraction motor casing length.

is postulated to range between 40% and 80% of the motor casing length but no basis for the assumed range is provided. The basis could be more qualitative than quantitative, but it still should be provided.

CNWRA34 Page 5-2, line 20 The FEDB is reference [35]. Reference citation corrected 10/14/19 NEI-33 5.1 The guidance and analysis presented and described in this Revised as suggested: 10/4/19 section apply only to the determination and analysis of electrical fires in electric motors (bins 2, 14, 21, 26, and 32). The guidance and analysis presented and described in this section apply only to the determination and analysis of electrical fires in Suggest adding Bin 9 (air compressors) as part of the electric motors (bins 2, 9, 14, 21, 26, and 32).

analysis for electric motor heat release rate and fire growth.

And revised throughout report.

NEI-34 5.1.5.1 (pg 104) Only a small fraction, 2 events, describe fires that caused Revised as: 10/4/19 extensive damage to the ignition source.

Only a small fraction, two events, describe fires that caused extensive damage to the ignition If the intent of this sentence is to convey "extensive damage source. Even with extensive damage to the ignition to the ignition source" as grounds for assuming threat of source, the external extent of damage is not likely external damage to targets outside the ignition source, it to be large.

should be clearly noted.

NEI-35 5.1.5.2 (pg 105) Figures 5-1, 5-2, and 5-3 Revised as: 10/4/19 Include the "<" symbol within the figures to denote less than the stated HP values on the axes. Include axis bars on figure 5-3.

Page 23 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date CNWRA35 Page 5-8, lines 21 and See general comment. (CNWRA33) Text revised as: 10/4/19 24 Based on the photographic reviews, the fraction

, ranging from 60-90% of the motor casing radius, is used to estimate the amount of cable material within a motor.

NEI-36 5.1.6.1 (pg 110) The wiring used in the rotor and stator are insulated with a Revised as suggested 10/4/19 coating that is assumed to be flammable. that is assumed to be flammable.

Typo - Remove second " that is assumed to be flammable."

NEI-37 5.1.6.1 (pg 110) Based on a review of pictures showing the internal Revised as: 10/4/19 components of electric motors, the cables within an electric Based on a review of pictures showing the internal motor do not run the entire length of the motor casing is components of electric motors, the cables within an assumed to contain the cables and conductors. electric motor do not appear to run the entire length of the motor casing (which is assumed to Sentence is unclear. contain the cables and conductors).

Page 24 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date CNWRA36 Revised as: 10/4/19 What does burn with a low flame heat flux mean. These These products are commonly made of halogen-Page 5-9, line 21 products are difficult to ignite and have a low heat release containing/"ame-retarded (FR) plastics, which are rate. difficult to ignite and have a low heat release rate.

CNWRA37 Revised as: 10/4/19 It is likely that only a fraction of the motor winding will ignite during any given event. In some instances, a small fraction may ignite and in others, the full area of the cylinder could be Page 5-9, line 27 See general comment. (CNWRA33) involved. To capture a range of likelihoods, an ignited perimeter and length fraction is applied to account for the variability in the ignited surface area of the cylinder. Given the low certainty of what fraction of the surface area would reliability be expected to ignite, these fractions are postulated to range between 25-100%.

CNWRA38 A negative heat release rate is not physical. It may be Footnote removed. 10/15/19 helpful to briefly explain how HRR0 is determined.

Essentially, specimens of a material are tested in the cone calorimeter or similar device to measure the heat release rate as a function of radiant heat flux. The measured HRRPUA (usually the peak value) is plotted versus heat flux, a straight line is fitted through the data points and Page 5-9, Footnote HRR0 is determined as the intercept of the linear fit with the ordinate (HRRPUA axis). Lyons group at the FAA has shown that polymeric materials with a HRR0 greater than 100 kW/m2 are unlikely to pass a small-scale flammability test such as the UL 94 vertical flame test. Materials with a negative or very low HRR0 require an external heat source to sustain flaming combustion.

CNWRA39 Page 5-10, line 5 See general comment. (CNWRA33) Revised as: 10/4/19 Page 25 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Therefore, a wire coverage fraction is applied to account for the available surface area for combustion. Given the significant fraction of wiring expected to be open and prone to igniting, this fraction is postulated to result in a range between 50-100% of the cable exposed.

CNWRA40 Yes. Revised as: 10/4/19 These values were determined through a uniform Page 5-10, line 10 Are all variables assumed to be uniformly distributed? randomization of all variable parameters discussed above, namely exterior motor dimensions, length fraction, radius fraction, HRRPUA, ignited area fraction, and wire coverage fraction.

NEI-38 Similar to the HRRs presented in NUREG/CR-6850 [1] and Revised as: 10/4/19 NUREG-2178 [2] utilize a gamma distribution is selected for Similar to the HRRs presented in NUREG/CR-two reasons: 6850 [1] and NUREG-2178 [2] a gamma 5.1.6.1 (pg 113)

Similar to the HRRs presented in NUREG/CR-6850 [1] and distribution is selected for two reasons.

NUREG-2178 [2], this method utilizes a gamma distribution which is selected for two reasons:

NEI-39 2. It produces a good fit to the data while maintaining a Revised as: 10/4/19 similar technical approach as previously used. 2. It produces a fit representative of the data 5.1.6.1 (pg 113)

Suggest more technical wording to replace "good fit" within while maintaining a similar technical approach as this justification previously used.

NEI-40 In some cases, these temperature ratings are greater than The intent of the sentence is to shape the 10/4/19 the steady state failure criteria for thermoplastic electrical discussion around the material properties used to cables, therefore it is considered appropriate to treat the estimate the HRRPUA used in the Monte Carlo 5.1.6.1 (pg. 111) insulation as having superior performance relative to a analysis. No judgment or/changes to the [damage]

thermoplastic cable. criteria of the motor cables is being suggested.

Revised to include:

Page 26 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date If the intent of this sentence is for the analyst to assume These enamel types include, but are not limited thermoset cable criteria, it should be clearly noted. to, thermoset type cable materials, No thermoplastic type cable material was used.

CNWRA41 Text revised as: 10/4/19 Using a detailed review of motor dimensions over various horsepower ratings a randomized How were the motor length and radius ranges determined sample size of 10,000 occurrences was Page 5-12, Table 5-2 performed and the distribution of fire size was for each rating determined?

estimated. From this distribution a corresponding gamma distribution fit to the resulting data. The results from this exercise are presented in Table 5-2.

CNWRA42 This and many other referenced information was 10/15/19 gathered by reviewing commercial specifications for motors. Later it was determined to not reference these documents as to avoid the appearance of commercial endorsements. Added Several of the characteristics of the wire and the Page 5-13, line 15 See general comment. (CNWRA33) stator windings were approximated based on expert judgement of images of exposed motor windings, motor specifications, and commercial product descriptions. Important characteristics were identified and assigned a randomized parameter appropriate to the observations.

CNWRA43 Page 5-13, line 22 See general comment. (CNWRA33) See response to comment CNWRA42. 10/15/19 CNWRA44 See response to comment CNWRA42. Revised 10/15/19 Page 5-13, line 31 See general comment. What is meant by observed?

observed to considered.

CNWRA45 Deleted the word observed and added 10/15/19 What is meant by observed? In this case the range was clarification. This was based on data found to Page 5-13, line 33 determined on heat of combustion data for the different correspond to the materials that were described.

jacketing materials that are used.

References are provided.

CNWRA46 Page 5-14, Table 5-4 See comment for Table 5-2. Table uses same groupings as Table 5-2. 10/4/19 Page 27 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date How were the growth, steady burning and decay durations Discussion and equations added to help clarify the CNWRA47 Page 5-15, lines 4-6 10/9/2019 determined? formulation of the results NEI-41 The feeder cable to the adjacent motor is routed directly Revised as: 10/9/2019 behind and adjacent to the ignition source. The feeder cable to the adjacent motor is routed directly behind and adjacent to the ignition source 5.1.7 (pg. 118)

It is unclear as to the purpose of this sentence. If the intent and was not damaged by the fire.

is to provide evidence that no external damage occurred during the described event, it should be clearly stated.

NEI-42 This is confirmed by investigation of Revised (VLO, 09/17/2019) 10/9/2019 5.1.7 (pg. 119) Stray paragraph break between lines 12 and 13 of this page.

NEI-43 5.1.8: "Growth: 2 minutes, t-squared growth" Section 9.3.1 revised to 2 minutes 10/9/2019 9.3.1: "Growth: 3 minutes, t-squared growth" 5.1.8 (pg 5-18) 9.3.1 (pg 9-2)

Fire growth duration of electric motors in Section 9.3.1 does not match what is provided in Section 5.1.8 NEI-44 The scope of this evaluation does not impact the treatment Revised as suggested. 10/9/2019 of fires in oil filed transformers which are expected to have substantially different burning characteristics due to the liquid fuel.

5.2 (pg 120)

The scope of this evaluation does not impact the treatment of fires in oil filled transformers which are expected to have substantially different burning characteristics due to the liquid fuel.

CNWRA48 Added clarification: 10/15/19 Several of the internal characteristics of the Page 5-25, lines 15 and transformer windings were approximated based on See general comment. (CNWRA33) 20 expert judgement of images of exposed transformer windings, specifications, and commercial product descriptions. Important Page 28 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date characteristics were identified and assigned a randomized parameter appropriate to the observations.

CNWRA49 Ultimately, the estimated HRR is tied to a single 10/9/19 cylinder in the Monte Carlo analysis. However, from the radius of the cylinder is estimated by the overall enclosure size.

Revised as:

This fraction, , is a range postulated between 50-80% of the widest dimension of the external transformer housing, ratioed with the smallest dimension. This method results in an estimation of What is the purpose of showing a single cylinder? It is the cylinder radius, r, that ensures the three Page 5-25, Figure 5-19 unlikely that L in this part of the figure is the widest cylinders defined will fit within the transformer dimension of the external transformer housing. enclosure.

Figure updated to revise L to H CNWRA50 Page 5-26, lines 21-22 See general comment. (CNWRA33) See response to comment CNWRA48. 10/15/19 CNWRA51 Text deleted and revised as: 10/9/19 Six out of twenty-eight (21.4%) outcomes that lead to The sizes for dry-transformer over a range of kVA Page 5-26, lines 25-26 ignition of all three cylinders does not seem to be consistent were reviewed to develop an appropriate and lines 40-42 with the statement in lines 25-26. estimation for HRR of a dry-transformer fire. The review of dry-type transformer fires from the FEDB Page 29 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date indicates that it is unlikely that the fire will consume a substantial portion of the contents of the entire transformer.

CNWRA52 No, Figure 5-20 is correct. 10/9/2019 CNWRA-52 Follow-up: Check whether references to Figures 5-19 and 5-20 on pages 5-26 and 5-27, Page 5-27, line 13 Should this refer to Figure 5-19?

respectively are correct.

Re-reviewed and these figure references were not corrected. Updated figure reference based on comment.

CNWRA53 Replace in each classification group with for each Revised as suggested. 10/9/2019 Page 5-28, line 17 power rating NEI-45 2. It produces a good fit to the data while maintaining a Revised similar to comment NEI-39 10/9/2019 similar technical approach as previously used.

5.2.6.1 (pg 130)

Suggest more technical wording to replace "good fit" within this justification CNWRA54 Text added after bulleted list: 10/9/2019 The height and radius ranges for various power How were the coil height and radius ranges determined for Page 5-29, Table 5-7 ratings were determined following a detailed each rating determined?

review of transformer dimensions CNWRA55 This sentence was trying to say that the best fit 109/2019 produces a distribution(s) nearly identical to the MC sampling. Text is deleted.

It is not clear which distributions are nearly Page 5-29, line 15 Referenced text:

indistinguishable.

The gamma distribution parameters are approximated from the sample size where Beta is determined from the variance of the sample set Page 30 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date divided by the mean of the sample set, then Alpha is the mean of the sample set divided by Beta.

Since the randomized set of 10,000 occurrences can vary, the precise evaluation of these terms can also vary, but the resulting distributions are nearly indistinguishable.

NEI-46 Any increase in the ZOI due to the increase in the HRR Revised as: 10/9/2019 following the guidance provided in Table 5-8 compared to Due to the increase in HRR, only a modest the peak HRR following the guidance provided in increase in the ZOI is expected and is likely to be NUREG/CR-6850 of 69 kW will be on the order of inches. bound by the ZOI of the associated switchgear Specifically with the increase from 69kW to 130 kW peak thermal fire.

5.2.6.1 (pg 132) HRR, the plume ZOI for thermoplastic targets increases by

~1.8 ft and the radiant ZOI increases by ~9 inches.

Likewise, the thermoset plume ZOI increases by ~ 1.4ft and the radiant ZOI increases by ~6 inches. Suggest clarification, plume ZOI changes by more than a matter of inches..

CNWRA56 Page 5-30, line 27 See general comment. (CNWRA33) See response to comment CNWRA48. 10/15/19 CNWRA57 Page 5-30, line 35 See general comment. (CNWRA33) See response to comment CNWRA48. 10/15/19 CNWRA58 Page 5-31, Table 5-9 Same comment as for Table 5-2. Table uses same groupings as Table 5-7. 10/9/2019 FJ-1 5.2.6.3 Fire base height. Does #1 and #2 conflict? If it is sealed we The rules do not appear to conflict. The second 10/9/2019 use 1 below top. If there is a low vent, we use the top? rule applies to open top transformers. The rule is revised as:

1. It is recommended that the assumed fire location (for screening and detailed fire modeling calculations) for dry-type transformers sealed on the top (without top vents or openings that allow vertical air flow) be 0.3 m (1 ft) below the top of the transformer. This assumption is also valid (conservative) for fully sealed transformer enclosures.

Page 31 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date

2. For a fire within a dry-type transformer that is not sealed at the top (i.e. openings located on the top surface of the transformer) the assumed fire location should be the top of the transformer.

The third bullet provides guidance for transformers with vents.

CNWRA59 Page 5-33, line 12 Replace the resulting maximum horizontal with Revised as suggested. 10/9/2019 the resulting maximum vertical CNWRA60 Page 5-35, line 6 Is the use of a fixed Froude number of 1.0 justified? The Monte Carlo results were used to determine 10/18/19 the range of Froude numbers associated with the new HRR classifications. A new discussion and table with classification specific values were added.

The examples were updated using the values presented in the table.

For comparison, the change in the critical HRR for a target located 0.9 m above a motor using the example Q* of 0.49 is 20.6. Using a Froude number of 1.0 results in a critical HRR of 17.2.

CNWRA61 Page 5-38, line 28 Same as previous comment: Similar to motors, examples revised to use Froude 10/18/19 Is the use of a fixed Froude number of 1.0 justified? number values added to the report, derived from the Monte Carlo analysis.

VLO 5.2.6, Figure 5-18 Suggest making a few more edits to the XFMR figure. This Re-created figure in PowerPoint 10/9/2019 comes from one of the references used to size the transformers for use in the Monte Carlo analysis. Its still very similar and should be further revised, or cited.

Section 6 (Fire Location Factor)

NEI-47 Up until this point in the report, fire diameter has been The equations presented in Chapter 6 come from 10/9/2019 6 calculated using the guidance found in Section 4.2 of the referenced EPRI report. The referenced report NUREG-2178 Volume 1 (i.e., enclosure footprint or preceded the publication of NUREG-2178, Vol. 1 Page 32 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date assigning the characteristic fire diameter). It is not clear how and the equations are included to provide equation 6-3 of this report was derived nor is it clear what background in to the methodology used.

the heat flux value represents. Suggest using the guidance The term should have been identified as the found in Section 4.2 of NUREG-2178 Volume 1 OR provide HRRPUA. Text revised to identify the HRRPUA.

reference for equation and define what the heat flux variable is meant to represent and if a radiative fraction needs to be included.

NEI-48 The location factor is typically not used for calculating the Prior to the presentation of the method described 10/9/2019 fire diameter. Remove location factor from fire diameter in NUREG-2178, Vol.1, for which a diameter could equation. If location factor is necessary, provide be estimated to ensure it remained within the justification. validation limits, the diameter could be estimated as a circular area given a specified HRRPUA.

The estimation of area would follow:

= 2 4

6 (pg. 6-2; Equation 6-

3) The would be used to estimate the area, A.

Solving this equation for D would be:

4

=

Since the HRR is now being used to estimate the diameter, and the location factor is included as a multiplicative term on the HRR to account for wall and corner effects, the location factor must be included in the estimated of the diameter.

CNWRA62 Add the following at the end of the sentence: and " is Revised as part of comment NEI-47 (VLO, 10/9/2019 Page 6-2, line 5 09/17/2019) the HRRPUA (kW/m2).

NEI-49 6 (pgs. 6-5, 6-6, and 6- "...different bias using both the traditional image method fire Revised to remove image 10/9/2019

7) location factor and the modified fire location factor.." Revised figure keys to use traditional.

Page 33 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Define traditional image method fire location factor or just state "traditional fire location factor". Update key in Figures 6-3 and 6-4 to include location factor in the description of each data set or figure titles to include mention of location factor.

NEI-50 "The temperature profile presented in Figure 6-1 remains For information and discussion on the NIST test 10/9/2019 primarily steady at various distances from a wall surface" data/results please see the referenced report.

6 (pg. 6-4, line 12) There is a dip in the temperature profile in Figure 6-1 and it is not addressed in the text. Include statement from NIST test which addresses this dip.

CNWRA63 Page 6-6, Figure 6-3 A possible explanation why the experimental values are When comparing the experimental results to the 10/9/2019 consistently lower than the model predictions is that the Heskestad equation, the NIST report assumed a radiant fraction was underestimated in the calculations (see radiant fraction of 0.25 for use with the natural gas comment on page 2-2, line 28). Moreover, the predictions burners. This is similar to the assumed value of 0.3 based on the image method are consistently higher than for used in the comparison presented in Figures 6-3 the FDS calculations because the former essentially and 6-4. Additionally, the radiant fraction values assumes that the walls are perfectly insulated while the associated with natural gas are usually less than latter takes heat losses through the walls into account. It is 0.3.

suggested to add these observations to the discussion of Figure 6-3 on page 6-5.

The original analysis documented in the referenced EPRI report considered sensitivity to wall surfaces.

The results were found to not be sensitive to wall type (including the inert type). See 3.1.4 in the referenced EPRI report.

No changes made.

CNWRA Follow-up Editorial (last sentence in the paragraph at the top of p. 8-4): Suggest changing the sentence to The new fire location factors provide a significant reduction in bias and model uncertainty over those recommended in Page 34 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date NUREG/CR-6850 [1] and the NRC Significance Determination Process [54], as shown in Table 6-2.

Response to follow-up: Accepted and made change in conclusions section.

CNWRA64 Page 6-7, Figure 6-4 Same as previous comment: Same as previous comment. 10/9/2019 A possible explanation why the experimental values are consistently lower than the model predictions is that the radiant fraction was underestimated in the calculations (see comment on page 2-2, line 28).

Moreover, the predictions based on the image method are consistently higher than for the FDS calculations because the former essentially assumes that the walls are perfectly insulated while the latter takes heat losses through the walls into account. It is suggested to add these observations to the discussion of Figure 6-3 on page 6-5.

Section 7 (NSP Floor) - also put into NUREG-2230 DH-1 NUREG-2178 Chapter 7 is very confusing as to what the was using older April version that disagreed. 10/9/2019 conclusion is providing. By reading it carefully, the chapter Comment no longer relevant in new version.

seems to propose a floor value for NSP of 1.0E-05.

However, Chapter 8 provides a recommended floor value of 7

1E-04 (Table 807). So 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.

GG-1 This section is intended to re-evaluate the floor value of the There is no double counting the frequency. The 10/9/2019 MCR non-suppression probability as a replacement of the value developed in this section is a 'floor' or limit.

Pg 7-1 Line 12-14 0.001 value that is currently used. It should be noted that This value will serve as the lowest value for which "the probability of having a fire at some point" in the MCR is a Pns value would be considered appropriate and Pg 7-2 Line 35-37 already covered by the fire ignition frequency. Therefore, acceptable for use in the Fire PRA. It is not used to including it in the non-suppression probability will result in double counting the frequency component.

Page 35 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date determine the scenario specific Pns or subsequent scenario frequency.

The introduction discussion has be simplified to remove references to the Pns model in NUREG-2230.

GG-2 While FAQ 12-0064 does allow for a zero (0) Hot 10/9/2019 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 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, 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 in NUREG-2230.

GG-3 The count is reflected in the ignition source 10/9/2019 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 Pg 7-1 Line 31 seemingly applies the full Bin 15 generic frequency to the the proposed floor value includes a single cabinet, 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 in NUREG-2230 to describe that the ignition source weighting factor is credited in the calculation, not the simply the generic frequency.

GG-4 10/9/2019 Saying the MCR represents 20% of the total floor area of Pg 7-2 Line 6 the control / auxiliary / reactor buildings is overly A sensitivity case has been added to Table 3-5 conservative. assuming the floor area ranges between 1% and 10% in NUREG-2230.

Page 36 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date GG-5 The commenter is correct that the mission time 10/9/2019 plays no direct role in the calculation of a MCR fire scenario frequency. It does, however, provide a period duration typically used in PRA analyses.

As noted in the draft NUREG, the goal is to estimate a floor value that represents the best possible manual PNS. This is done considering both the probability of having a fire and failing to suppress that fire.

This consideration is developed under the idea that an unsuppressed Control Room fire within a period The statement that "the Fire PRA estimates the probability of hours is very low likelihood. This is consistent of an event occurring with the OPEX, which so far suggests fire as over a 24-hour mission time" does not make sense in the suppressed in less than 10 minutes given the context of fire scenario continuous pretense of operators in the room.

frequency calculation. The plant mission time in PRA refers Since the floor is a value set up to prevent Pg 7-2 Line 11-14 to the time needed for unrealistically low estimations of the PNS, a value the mitigation systems to bring the plant in a safe long-term 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 condition. Therefore, the mission time plays no role in the calculation of the MCR fire scenario practical limit (i.e. this floor help to prevent a PNS from being unrealistically low as it will be higher frequencies. 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 Page 37 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date 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 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.

GG-6 The calculation of the non-suppression probability floor The comment is incorrect that the proposed floor in 10/9/2019 value presented in Section 7.3 does not appear to have a the draft NUREG would result in a scenario sound basis. If applied in practice as a direct frequency equation similar to: Ignition Frequency Pg 7-2 Line 35-37 replacement of the 0.001 floor NSP, this approach would x...x (Ignition Frequency expressed as a probability result in the following scenario frequency equation :

for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />) x (Average NSP).

Scenario Frequency = Ignition Frequency x...x (Ignition Frequency expressed as a Page 38 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 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 in NUREG-2230 revised to number unrealistically low.

clarify that the development of the stated floor value is a calculation of a scenario PNS value.

NEI-51 The first step uses the MCR suppression rate in Table 8-6 The discussion on the NSP floor was moved for calculations of the Pns with a floor value of 1E-03. This into NUREG-2230.

results in fires that are suppressed prior to approximately 18 Per the text in Section 3.5.2:

minutes using a suppression rate of 0.385. The second step captures all remaining MCR fire durations up to the The second step captures all remaining MCR fire proposed floor of 2.4E-07 by making use of the ignition durations up to the floor of 2.4E-07 through the use 7.4 source bin specific suppression rate. For example, a fire in a of an ignition source bin specific suppression rate (7-3) cabinet located within the MCR will use the Interruptible and (interruptible, growing, transient, etc.).

Growing suppression rates presented in NUREG-2230 [62] No changes made.

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 NEI-67 Section 7.4, Figure 7-1, The first step uses the MCR suppression rate in Table 8-6 The discussion on the NSP floor was moved / 10/4/19 and Table 7-2 for calculations of the Pns with a floor value of 1E-03. This resolved into NUREG-2230.

(7-3, 7-4, and 7-5) 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 Equations for the calculation of the numerical proposed floor of 2.4E-07 by making use of the ignition suppression results for electrical cabinet source bin specific suppression rate. For example, a fire in a (interruptible and growing) and transient ignition cabinet located within the MCR will use the Interruptible and sources have been added.

Growing suppression rates presented in NUREG-2230 [62] (As of 9/12/19 needs to be updated using algebraic for fire durations in excess of 18 minutes. method as opposed to current ratio fitting.)

Provide equations so the values in Table 7-2 can be replicated. Statement is unclear how to apply the two-step calculation Page 39 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date CNWRA65 Duplicate to 2230 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 The discussion on the NSP floor was moved / 9/11/19 Page 7-1, line 18 the end of the paragraph. resolved into NUREG-2230.

CNWRA66 Page 7-1, line 28 For the MCB frequency, refer to Table 8-3 instead of The discussion on the NSP floor was moved / 9/11/19 Chapter 8. resolved into NUREG-2230.

CNWRA67 Page 7-1, line 33 For the electrical cabinet frequency, reference is made to The discussion on the NSP floor was moved / 9/11/19 NUREG-2230. The citation can be more specific (the value resolved into NUREG-2230.

is provided in Table 3-9 of NUREG-2230).

CNWRA68 Duplicate to 2230 Provide a basis for the 300-700 range of the number of The discussion on the NSP floor was moved / 9/11/19 Page 7-1, lines 35 cabinets counted as ignition sources in a single unit NPP. resolved into NUREG-2230.

This seems inconsistent with the average plant-wide count of 750 in Table 6.2.3 of IMC 0308 Attachment 3 Appendix F

[54].

CNWRA69 Duplicate to 2230 Provide a basis for the 10%-30% range of the apportioning The discussion on the NSP floor was moved / 9/11/19 Page 7-2, lines 4-6 factor (fraction of the control, auxiliary and reactor building resolved into NUREG-2230.

floor area representing the MCR).

CNWRA70 Duplicate to 2230 Delete the sentence This results 1.21E-4/day. because The discussion on the NSP floor was moved / 9/11/19 Page 7-2, lines 12-14 this calculation is incorrect, as it does not account for the resolved into NUREG-2230.

and Equation 7-1 random variables (total number of cabinets in the plant and apportioning of the area of the MCR). Instead, change Equation 7-1 as follows:

Pr(t24 h)=((3.7E 03))(365) 1E 05 CNWRA71 Duplicate to 2230 According to NUREG-2178, Volume 1, the 98th percentile The discussion on the NSP floor was moved / 9/11/19 Page 7-2, line 26 peak HRR of a large open electrical enclosure with resolved into NUREG-2230.

thermoplastic cable contents is 1,000 kW. Five minutes after ignition the fire in such a cabinet would grow to 174 kW. Three minutes after ignition the HRR is 63 kW. A value closer to 3 minutes may be a more appropriate lower limit for the range.

Page 40 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date CNWRA72 Duplicate to 2230 Explain how the non-suppression floor value for a dual unit The discussion on the NSP floor was moved / 9/11/19 Page 7-2, line 41 MCR was estimated. resolved into NUREG-2230.

Section 8 (Main Control Board)

CNWRA73 The two characteristics are described in the 10/15/19 preceding sentence. A small fire with a small What are the two characteristics that are being referred to in Page 8-1, line 16 damage consequence, and suppression prior to the sentence that starts with These two characteristics substantial growth or propagation. Will revise the sentence to clarify the discussion.

GG-7 This report does not define screening values, and 10/15/19 users should follow existing guidance for screening This section does not suggest a threshold CDF value for values. The following clarification was added.

Page 8-3 Line 30-32 screening purposes. It is not clear how low the CDF should This report does not propose specific screening be so that it can be considered sufficiently low. values, and analysts should follow existing guidance in the selection of an appropriate screening value.

GG-8 This section tries to describe how to estimate the CCDP for The CCDP is conditional on the failure of the 10/17/19 when fire damage is limited to a single subcomponent on a component (i.e. the probability of failure is 1.0). It MCB panel. It is stated the risk contribution from such a fire inherently does not include ignition/failure should be bounded by "the quantification of the internal frequency. Similarly, the plant response model is events model as random failure probabilities including those supposed to capture the effects of a single associated with ignition of individual subcomponents component failure causing failure of subsequent credited in the model". It is further stated "...the model components. These factors should be captured if quantification assuming a plant trip from the internal events the metric of comparison is CCDP. Added Page 8-18 model with no fire induced failures can provide a bounding clarification that the quantity of interest is CCDP, CCDP value to use for calculating the CDF associated with and the following statement:

this damage state".

It needs to be noted that the statement in this paragraph is Note that the internal events model should correct only if the consider plant responses such as subsequent following two conditions are met: spurious operations of equipment caused by the

1) The ignition probability is included in the components failure of the initial subcomponent.

random failure rate, i.e. it is not counted out because it is Page 41 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date supposed to be covered by fire ignition frequency in Fire PRPA.

2) The component failure does not cause another more onerous initiating event. For example, a spuriously open pressurizer PORV, spuriously open steam dump etc. may be triggered by MCB single component failures.

GG-9 Clarification should be given on what qualifies as an internal Provided reference to Chapter 4.2 for screening 10/15/19 Page 8-22 Line 24-25 partition that would prevent fire spread to adjacent panel. criteria.

GG-10 Clarification is needed on what constitutes the total surface Guidance is provided. 10/18/19 area of the panel. For some panels, a greater proportion of the surface area is simply empty with no Page 8-22 Line 27-28 controls / indication. If this empty surface area is accounted for in the apportioning of the frequency then it would effectively dilute the frequency assigned to the panel based on linear width.

GG-11 General point, this circle ZOI approach applied to the The idea here is that the primary ignition event is a 10/18/19 surface of panel does not account for any cable grouping or component, and not necessarily a cable bundle.

cable bundles which could be present Added footnote:

Section 8.5 - All immediately behind the panel surface. If is often seen that cables are bundled together and fire damage to these This approach is preferred over a fixed grid bundles would expand the damage beyond the approach as a fixed grid can miss important combinations of targets based on the arbitrary 1.0ft2 ZOI range. location of the grid lines.

GG-12 Corrected, though technically all branches from B 10/15/19 onward are dependent on the parameter. Clarified Page 8-23 Line 7 Should refer to Branch B (not Branch C). as follows:

(i.e., Branch B in the event tree corresponding to the successful suppression split)

CNWRA74 The radiant fraction is assumed to be 0.3. As discussed This comment was also made in the draft report for 10/15/19 before, this value may be too low for MCB fires. NUREG-2230. Since both methods rely on a Pages 8-23 and 8-24, similar monte-carlo approach, the two methods Table 8-10 The HRRPUA is assumed to range between 150 and 500 have been revised to incorporate similar changes.

kW/m2. The lower limit is the generic value for TP cable but 1. Radiant fraction, soot yield and heat of no basis is provided for the upper limit. combustion are now a randomized Page 42 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date parameter based on fuel property selection for TP/TS cables listed in the SFPE Handbook, 5th Ed, Table A.39. Only cables with all three parameters defined are selected in the randomization.

2. HRRPUA value ranges have been revised to match the recommended ranges provided in NUREG-7010 for TS/TP cables. 100-200 and 200-300 kW/m2 respectively.

3. Split fraction for interruptible/growth fires and associated NSP calculations Details of these changes have been provided. The tables in Section 7.5 and onwards have been updated to reflect this change. Tables that were updated include: Table 7-10, 7-12, 7-13, 7-14, 7-

15. The example was also updated including Tables 7-18 and Table 7-19.

GG-13 The value 0.15 m (6 in) is the intended distance 10/17/19 since it corresponds to the IEEE standard. Revised The radial distance of 0.15m (6in) seems confusing. Should the sentence on page 8-22 to clarify:

Page 8-12 Line 13 it not correspond to the 0.17m (7in) defined on page This corresponds to roughly 17 cm (7 in) from the 8-22, line 10? point of ignition in any direction if considering a circular ZOI or roughly 15 cm (6 in) if considering a square ZOI.

GG-14 Page 8-31 Line 2 The NSP floor value should read 2.4E-07 (not 2.7E-04). Revised as suggested. 10/15/19 GG-15 "..are set to a value of one..." needs to be changed to "..are Revised as suggested. 10/17/19 Page 8-34 Line 1 set to zero...".

GG-16 It is not clear why it has been recommended a single The average value was intended as an alternative 10/16/19 averaged value to be applied, when the spread fraction to the table for simplicity. Removed Page 8-34 Line 4-6 associated with the specific MCB configuration can just as recommendation, instead analyst will use the easily be applied.

Page 43 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date values from the table that correspond to the specific ignition source.

NEI-52 This is an expected trend in the value of 1-delta. 10/15/19 Since the non-suppression term in Table 8-13 are smaller than in Table 8-12, the fires that remain

(>98th percentile) after the split are a higher percentage. The trend in the values is dependent on the previous values. Added the following clarification:

In Table 8-15 there are values for in-cabinet detection is credited that are worse than values in Table 8-14 where Note also that some values in Table 8 15 for 8.6 (pg 8-34) cabinets with detection are higher probabilities there is no in-cabinet detection credited for the same enclosure type, which seems counterintuitive. than in Table 8 14 for cabinets without detection due to the previous credits from Table 8 12 and Table 8 13. The more suppression credit that occurs in the prior branch, the less credit that can occur in the subsequent branch since the remaining fires will have a higher percentage of high severity. The overall result, considering the product of the two values, is lower risk when cabinets are equipped with in-cabinet detection.

NEI-53 The data in Tables 8-14 and 8-15 appears to suggest that See response to previous comment. Note that the 10/15/19 the Fraction of Fires that Spread to an Adjacent Panel difference between TP and TS cables overall is

() is more severe for Thermoset cabinets than smaller due to the HRRPUA and its associated Thermoplastic cabinets, (e.g., 0.114 (TS) vs 0.078 (TP) in impact on the fire diameter. Since the TS cable has 8.6 (pg 8-34)

Table 8-14 for 4a-Closed MCBs). This seems a lower HRRPUA, it will have a correspondingly counterintuitive given that NUREG/CR-6850 and NUREG- larger diameter and therefore damage cables due 2178 Volume 1 Suggest Thermoplastic fire HRRs/Severity to flame spread faster.

factors are more severe than those of thermoset fires.

NEI-54 8.9 The manual suppression rate is defined as 0.385 min-1 from The reference to the control room suppression rate 10/17/19 (8-39) Table 8-7. has been corrected to Table 7-6.

The manual suppression rate is mentioned in Table 8-6 Page 44 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date NEI-60 Page 8-16 Given the new NSP floor described in Section 7, the The analysis and discussion have been revised to 10/17/19 numerical results for the Control Room suppression curve account for growth/interruptible split. This change are presented in Table 8-7. was made to be consistent with NUREG-2230 It is unclear how this should be applied relative to the which uses a similar monte carlo approach.

growing/interruptible fires described in section 7.4. If this version is strictly for Bin 4, this should be noted.

NEI-61 Based on this ZOI, the recommended process consists of A fixed grid approach may miss some critical 10/17/29 identifying targets within circles of approximately 0.09 m2 targets that are within 1 ft of linear separation. The (1.0 ft2) throughout the surface of the panel. arbitrary location of the grid lines would then provide a false indication of the overall risk. The Page 8-22 judgement of the working group was to prefer the It would seem more practical for application of this guidance wandering circle/square approach and to avoid the to allow for a grid or square approach of the same size to be grid approach.

used to ensure all area of the board is covered and simplify the process. See prior response to comment GG-11.

NEI-72 0.3 lTypical fire radiant fraction The fixed radiant fraction approach has been 10/16/19 Since this an MCB specific analysis, reference to the revised to a variable approach consistent with Table 8-10 NUREG-2230.

radiative fraction of typical cable materials should be made to ensure this radiant fraction is acceptable. See prior response to comment CNWRA74.

NEI-62 recall that small fires under 20 kW are explicitly modeled in This sentence was deleted. This was an earlier 10/16/19 earlier branches of the scenario progression event tree assumption, but no longer necessary. Under the final approach, even fires less than 20kW have the Page 8-24 opportunity to damage nearby targets (this is now 20 kW is never explicitly stated as the limit in the discussion dependent on the scenario characteristics and of previous branches. If this is an equivalent based on the calculations in the Monte Carlo simulations).

1.0 ft2 ZOI, the inputs for this should be given such as TP or TS cable assumed, radiative or plume target assumed, etc.

CNWRA75 Provide references and include citations of the reports that References added. 10/17/19 Pages 8-24 through 8-describe the tests for which selected results are presented 29, Table 8-11 in the table.

Page 45 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date NEI-63 Calculation of fire ignition frequencies for each panel within Yes, the weighting factor is still applicable to this 10/17/19 the MCB. The ignition frequency for this screening process analysis. Added the following clarification.

is a multiplication of Page 8-3 with applicable weighting factors as defined in Does this include the location weighting factor from 6850? NUREG/CR-6850 (Ref. 1) i.e. factor of 2 for a 2 unit plant with 1 MCR.

NEI-64 8.3.2.2 Frequency Apportionment Example 2 Requested, but not necessary. Will input if 10/17/19 For this example, since the rear MCB panels are counted available.

Pages 8 8-14 separately, it would be useful to have an image of what the rear side looks like to illustrate the point. The current images in Figure 8-6 and 8-7 are only front side images.

CNWRA76 Pages 8-24 through 8- Provide references and include citations of the reports that References added. 10/16/19 29, Table 8-11 describe the tests for which selected results are presented in the table.

CNWRA77 Page 8-30, line 18 Exactly which guidance in Section 3 is being referred to? Clarified to section 2.2.3 for unobstructed radiation. 10/16/19 CNWRA78 Page 8-31, line 2 For the floor value refer to Table 7-1 or include a citation of Added reference. 10/16/19 NUREG-2230.

CNWRA79 Page 8-31, line 10 For the updated manual suppression rate refer to Section Included reference to Table 7-6. 10/16/19 8.3.3 or Table 8-6.

CNWRA80 Page 8-33, line 3 Replace effect with affect Replaced. 10/16/19 CNWRA81 Page 8-34, lines 1-3 Add the underlined words in the first line: The values of Added of (1-)

for Very Low fuel loading CNWRA82 Page 8-34, lines 4-6 Assuming a single value of 0.126 will result in The average value was intended as an alternative 10/16/19 underestimating the probability for spread for 11 of the 32 to the table for simplicity. Removed cases. Is that acceptable? recommendation.

See response to GG-16, repeated below:

The average value was intended as an alternative to the table for simplicity. Removed recommendation, instead analyst will use the Page 46 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date values from the table that correspond to the specific ignition source.

CNWRA83 Page 8-39, lines 5-7 The table numbers are incorrect. They should be 8-6, 8-18, Agree 10/17/19 and 8-19 instead of 8-7, 8-15, and 8-18, respectively. Should be Table 7-6 (sup rate)

Table 8-18 in the example of input parameters Table 8-19 is the results.

CNWRA84 Page 8-40, Table 8-18 Replace for enclosures with no in-cabinet detection. Replaced. 10/17/19 with for large closed enclosures with default TP fuel loading and no in-cabinet detection. in the description for Pns(t1)

NEI-83 Table 8-17 and 8-18 Manual suppression rate constant from Section 8.3.2 Revised to refer to NUREG-2230 since additional 10/17/19 Update statement to reference Section 8.3.3 changes to the PNS term have been made.

Section 9 (Summary)

CNWRA85 Replace and fire location factor with and fire base Revised as suggested 10/9/19 Page 9-1, lines 3-5 elevation CNWRA86 The growth, steady burning and decays times should be 2, Revised 10/9/19 Page 9-2, lines 1-4 13, and 2 minutes, respectively.

CNWRA87 Replace the first sentence (starts with Section 6 and Revised as suggested. 10/9/19 ends with and corners. with the following two sentences: Section 6 summarizes the latest guidance for determining plume temperatures for fires postulated in a Page 9-3, lines 33-35 corner or along a wall. The section provides validation of the new fire location factors based on the National Institute of Standards and Technology (NIST) experiments near walls and corners.

NEI-55 9.3.1 The following growth, steady burning, and decay durations Revise to match Section 5.1.6.2 10/9/19 (9-2) should be used for motor fires.

Growth: 3 minutes, t-squared growth Page 47 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Steady burning: 20 minutes Decay: 3 minutes, linear decay Update statement to match Section 5.1.6.2, which states the fire growth for electric motors is:

Growth: 2 minutes, t-squared growth Steady burning: 13 minutes Decay: 2 minutes, linear decay CNWRA88 Delete the following sentence: Definitions of wall and Revise to Section 6.2 10/9/19 Page 9-4, line 2 corner configurations are presented in Section 6.1. There are no such definitions in Section 1.

NEI-56 Section 7 discusses the background on the NSP floor and The discussion on the NSP floor was moved into 10/9/19 provides a basis for lowering the floor to 2.4E-7 for fire NUREG-2230.

scenarios in single unit MCRs. A revised floor value for dual Section deleted.

9.5 unit MCRs of 3.5E-07 is also recommended based on a (9-4) sensitivity analysis for multi-unit control rooms.

Add statement discussing two-step method for calculation Two stage MCR suppression curve now in manual non-suppression for non-MCB fires over 18 minutes NUREG-2230.

in the MCR Appendix A (Heat Soak)

NEI-57 Appendix A Tables A-3, A-4, and A-5 refer to NUREG/CR-6805. Fixed. 10/3/19 All references to NUREG/CR-6805 should be changed to 6850.

BJ-4 Page A-4, 1st line Typo in the unit of Stefan-Boltzmann constant: kW/m²-K4 is Fixed. 10/3/19 correct.

CNWRA89 Page A-1, line 17 Replace inner jacket temperature with jacket Fixed. 10/3/19 inner temperature .

Page 48 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date CNWRA90 Page A-1, lines 23-24 Replace without the data processing and modeling time Fixed. 10/3/19 requirements with without the time consuming input data gathering and output data processing requirements CNWRA91 Page A-1, line 31 Replace report with Appendix Make this Fixed. 10/3/19 change throughout the Appendix.

CNWRA92 Page A-2, line 11 Replace are discussed methodology with are Revised pointer to specific section. 10/16/19 discussed in the section describing the methodology CNWRA93 Page A-11, line 5 Add the following (underlined part) at the end of the Fixed 10/3/19 sentence: (sizes and jacket and insulator materials) exposed to a constant radiant heat flux.

CNWRA94 Page A-11, line 6 Add the following (underlined part) to the words in Fixed 10/3/19 parentheses: (model predicts faster time to failure).

CNWRA Editorial Comments Page vii, line 15 Replace group composed with group was Revised as suggested 10/9/19 composed Page viii, line 8 Replace expertise implementing the methods with Revised as suggested 10/9/19 expertise from implementing and supplementing the existing methods Page viii, line 23 Delete are discussed. Revised as suggested 10/9/19 Page 1-1, line 28 Delete , including. Revised as suggested 10/9/19 Page 1-2, line 4 Replace NUREG/CR-6850 with NUREG/CR- Revised as suggested 6850, (comma added).

Note: This is an example of a reference to NUREG/CR-6850, Vol. 2 where the volume is not specified (see general editorial comment above)

Page 1-3, line 31 Replace presented including with presented, Revised as suggested 10/9/19 including Page 49 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Page 1-3, lines 35-36 Replace on implementing the results and insights of the Revised as suggested 10/9/19 Fire Dynamics (FDS) simulation on with for implementing the results and insights of the Fire Dynamics (FDS) calculations of Page 1-4, line 6 Replace is a consideration with need to be Revised as suggested 10/9/19 considered Page 1-4, line 7 Replace in a corner location and when with in a Revised as suggested 10/9/19 corner, and to determine when Page 2-1, line 11 Replace use of field model with use of a field Fixed in NEI-17 10/3/19 model Page 2-1, lines 24-25 This sentence seems out of place and can/should be Deleted. 10/3/19 deleted.

Page 2-2, line 3 It is custom to use Q for heat release rate and q for heat Fixed. 10/3/19 flow (and therefore q ^" for heat flux).

Page 2-7, line 18 Replace representative of a large with Fixed. 10/3/19 representative of large Page 2-8, line 11 Replace are large fires. with are larger. Fixed. 10/3/19 Page 2-12, line 12 Add volume and dimensions in SI units, i.e., 1.78 m3 (0.91 Fixed 10/3/19 x 0.91 x 2.13 m).

Page 2-14, line 6 Replace inside the cabinet with the door open. with Fixed 10/3/19 inside the cabinet in the tests with the door open.

Page 2-17, line 10 Replace high values resulting in unrealistic Fixed 10/3/19 (conservative) target damage. with high values and, therefore, in unrealistic (conservative) target damage estimates.

Page 2-19, line 6 Add the following sentence at the end of the paragraph: An Fixed. 10/3/19 example of the latter is shown in Figure 2-16.

Page 50 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Page 2-20, lines 12-15 Strictly speaking, a 50 ft3 is not a large cabinet. Is there a Yes, strictly speaking 2178 says <= 50 ft3 is 10/3/19 difference between m50 and l50? In addition, the unit medium; however, also strictly speaking conversions are soft conversions. Which units are exact? 50.000000000000001 ft3 would be large and the reality is from any modeling point of view this can be treated as a 50 ft3 cabinet. The difference between large and medium is the fire. Large cabinets have a different HRR distribution than medium so a 75th percentile l50 is per 2178 a different fire size than a 75th percentile m50.

Page 2-21, line 6 Where is the SE corner? If directions are important, it would Report text prior to the bulleted list states: Note 10/3/19 be helpful to show which direction is north in Figure 2-17. that in discussion of results, the face with the door or louvers is referred to as the north face, and its opposite face is referred to as the south face. Have added this to caption for 2-17 and again before figure 2-18 Page 2-22, lines 5, 13- Same as previous comment. 10/3/19 See above 14, and 17 Page 2-25, line 17 Delete that. Fixed. 10/3/19 Page 3-2, line 6 Replace Obs_Fac are with Obs_Fac values are Fixed. 10/3/19 Page 3-13, line 24 Replace footprint (the with footprint (they The parentheses does not read well with they. 10/3/19 Kept the range Page 3-15, line 2 Replace ORIENATION with ORIENTATION Fixed. 10/3/19 Page 3-15, line 31 Replace the fuel was with if the chemical formula Fixed. 10/3/19 of the fuel was specified as Page 4-2, line 10 Replace in fire protection/fire modeling literature. with Revised as suggested 10/9/19 in the fire protection/fire modeling literature.

Page 51 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Page 4-2, line 13 Delete and examples because the examples are in Revised as suggested 10/9/19 Appendix D as stated in line 21.

Page 4-2, line 31 Replace methodology require with methodology Revised as suggested 10/9/19 requires Page 4-5, lines 15-16 Replace both the NRC-RES/SNL [30] tests and the Revised as suggested 10/9/19 NUREG/CR-7197 fire tests with both the NRC-RES/SNL [30] and NUREG/CR-7197 fire tests Page 4-14, line 47 Replace as necessary to the analysis goals with Revised as suggested 10/9/19 as necessary to meeting the analysis goals Page 5-1, line 25 Replace Examples applications with Example Revised as suggested 10/9/19 applications Page 5-7, line 32 Replace Supplement 1 states with Supplement 1 Revised as suggested 10/9/19 to NUREG/CR-6850 states Page 5-8, line 2 Delete that is assumed to be flammable. at the end of the Revised as suggested 10/9/19 sentence.

Page 5-8, line 18 Replace motor casing is assumed with motor Revised as part of an earlier comment 10/9/19 casing that is assumed Page 5-8, line 26 Replace estimation pairs with estimates Revised as suggested 10/9/19 Page 5-14, line 13 Delete well with the NUREG/CR-7010 [24] calculations Revised as suggested 10/9/19 because it is redundant with the rest of the sentence.

Page 5-14, line 14 Replace method consistently with method, the Kept original wording 10/9/19 latter consistently Page 5-14, line 15 Replace provides rationale with provides a Revised as suggested 10/9/19 rationale Page 5-17, lines 2-3 Replace the sentence with The maximum HRR can be Revised as suggested 10/9/19 estimated using the dimensions listed above, the Solid Page 52 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Flame II radiation model, and Heskestads fire plume model

[3,38].

Page 5-17, line 12 Remove hard return. Revised as suggested 10/9/19 Page 5-19, line 6 Replace assign transformer with assign a Revised as suggested 10/9/19 transformer Page 5-20, line 19 Spell out BWR or use the acronym for both types of Both terms previously identified, revised to only 10/9/19 reactors. include acronym.

Page 5-23, line 5 Make Dry-Transformer plural. Revised as suggested 10/9/19 Page 5-28, line 2 There does not appear to be a reference to this figure in the Reference added 10/9/19 text.

Page 6-1, line 19 Replace corners or wall surfaces with a corner or Revised as suggested 10/9/19 a wall surface Page 6-1, lines 19-22 Replace the sentence The fire location factor along Revised as suggested 10/9/19 walls. with The fire location factor, kF, is used in calculating plume temperatures for an ignition source fire that is located in a corner or along a wall. Its effect is to increase the plume temperature at a specified elevation compared to that for the same ignition source fire in the open.

Page 6-3, line 43 Replace corners or along walls with a corner or Revised as suggested 10/9/19 along a wall Page 7-2, lines 26-27 Replace occurs before a half of the Growing with Section removed from report 10/9/19 occurs before half of the Growing Page 7-4, Figure 7-1 Growth Fire is referred to as Growing Fire in other Section removed from report 10/9/19 places. Strictly speaking, the lower limit for the y-axis is 0.00001. It would be better to print the axis labels and titles and the text in the legend in black.

Page 53 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date Page 8-7, lines 24 and Replace a MBC with an MBC Revised to a MCB (not MCB) 10/9/19 26 Page 8-7, line 33 Replace were with are Revised sentence and correction no longer needed 10/9/19 Page 8-10, Figure 8-2 The text is hard to read. The font size should be large and Enlarged as much as possible. 10/9/19 the text should be printed in black.

Page 8-11, Figure 8-4 Same as previous comment. Enlarged as much as possible. 10/9/19 and Page 8-12, Figure 8-5 Page 8-13, line 10 Replace liner with linear Corrected 10/9/19 Page 8-16, Figure 8-9 Axis labels and titles and legend text should be printed in Changed axis and title to black. Will print in color. 10/9/19 black.

Page 8-23, Table 8-10 The gamma distribution parameter in the heading of the Changed 10/18/19 fourth column should be . In the notes column, the reference should be corrected or added for the MCR Suppression Rate (Section 8.3.3 or Table 8-6) and the NSP floor value (Table 7-1 or NUREG 2230).

Page 9-1, line 30 Delete air gap. Revised as suggested 10/9/19 Page 9-1, line 35 Replace fuel (exposing) with fuel loading Revised to include load to match similar bullet text 10/9/2019 (exposing)

Page 9-1, lines 33 and Make items plural. Revised as suggested 10/9/2019 36-37 Page 9-2, line 14 Should refer to Table 9-1 instead of 8-1. Revised (but actually now 8-1 since NSP floor 10/15/19 section removed)

Page 9-4, line 2 Replace factor provides with factors provide Revised as suggested 10/9/19 Page 9-4, line 3 Insert over those recommended in NUREG/CR-6850 [1] The basis for the traditional fire location factor 10/9/19 and the NRC Significance Determination Process [54] after values does not originate in NUREG/CR-6850 or uncertainty. the NRC SDP and are generic values used across fire protection.

Page 54 of 55

REVIEW / COMMENT DOCUMENTATION Document #, Rev: NUREG-2178 Vol 2 / EPRI 3002016052 Date: August 2019 Reviewer Document Number Comment No. Review Comments (Print)/Basis for Comment Comment Disposition / Resolution Acceptance /

Section / Paragraph Date No changes made.

Page 9-4, line 8 Replace floor with floor value (2x) Section removed 10/9/19 Page A-1, line 11 Delete and. Fixed. 10/3/19 Page A-1, line 15 Insert Supplement 1 to before NUREG-1805. Fixed 10/3/19 Page A-11, line 8 Replace cables tests with cable tests Fixed. 10/3/19 Appendix D: General There are some inconsistencies in the capitalization of Section revised. 10/9/2019 certain words (enclosure, exposed, exposing, etc.) For Capitalization of the term enclosure is limited to example, in some cases Enclosure is used but sometimes identifying the cabinet of interest (ex. Enclosure 2),

enclosure is used. spreading rules (ex. Group 4c: Small Enclosure) ,

and NUREG-2178 functional classes (ex. Group 4b Enclosure Class)

Page D-1, Figure D-2 This figure also shows the HRR profile for fires that spread Revised as (VLO 09/25/2019): 10/9/19 from enclosure 5 to enclosure 4 and from enclosure 5 to Enclosures 4 to 5, 5 to 4, and 5 to 6: Enclosure to enclosure 6. Enclosure Fire Spread HRR Profile Page D-7, line 7 Replace Figure D-4 and D-5. with Figure D-4, D-5, Revised. 10/9/19 and D-6.

Page D-14, line 11 Closing parentheses are missing. Section significantly revised to follow NUREG-2230 10/9/19 event tree. Comment no longer relevant.

Page 55 of 55