ML20216C850
| ML20216C850 | |
| Person / Time | |
|---|---|
| Site: | Summer  |
| Issue date: | 04/09/1998 |
| From: | Padovan L NRC (Affiliation Not Assigned) |
| To: | Gabe Taylor SOUTH CAROLINA ELECTRIC & GAS CO. |
| References | |
| GL-88-20, TAC-M83680, NUDOCS 9804150072 | |
| Download: ML20216C850 (10) | |
Text
r April 9, 1998 X
l Mr. Gary J. Taylor Vice President, Nuclear Opera'.ons South Carolina Electric & Gr, Company i
Virgil C. Summer Nuclear Station Post Office Box 88 Jenkinsville, South Carolina 29065 SUBJECT-REQUEST FOR ADDITIONAL INFORMATION - GENERIC LETTER 80-20, INDIVIDUAL PLANT EXAMINATION OF EXTERNAL EVENTS (IPEEE) t L
SUBMITTAL (TAC NO. M83680) l
Dear Mr. Taylor:
l We need additional information from South Carolina Electric & Gas Company in order for us to
[
complete our review of your IPEEE submittal. We, with assistance of our contractors,.
Brookhaven National Laboratory and Sandia National Laboratories have developed the enclosed request for additional information (RAI). The RAI pertains to your IPEEE seismic and fire analyses. There are no RAls in the high winds, floods, and other external events area.
l Please provide your response by September 1,1998, to meet our review schedule. You may call me at (301) 415-1423 if you have any questions.
l Sincerely,
.(Original Signed By).
l l
L. Mark Padovan, Project Manager 1
Project Directorate _II-1 Division of Reactor Projects'- 1/ll Office of Nuclear Reactor Regulation I.
Docket No. 50-395 5
l
Enclosure:
Request for l
Additional Information ff* W w &&w.a [vA I99 5I5 cc w/ encl: See next page u
\\fl Distribution
.a W His-J. Zwolinski OGC l
,O) l PUBLIC L. Plisco, Ril ACRS pr j.
' PD ll-1 Reading A. Rubin E. Chelliah DOCUMENT NAME: G:\\ SUMMER \\ SUM 83680.RAI Ya receive a cepy of this document. indicate in the bos: *C" = Copy without attachment / enclosure
'E" = Copy with attachment / enclosure "N" - No copy OFFICE PDIl 1/PM
.o lE) PDII 1/LA lE l
PD!!-1/C/)f L l
NAME MPadovan M
EDunnington ffr)
P.T. Ku?
DATE 04/ 9 /98 04/ 9 /98 04/O/98 0FFICIAL RECORD COPY l
l l
Q-)GyQ 9804150072 980409 PDR ADOCK 05000395 p
o
[
ENCLOSURE REQUEST FOR ADDITIONAL INFORMATION VIRGlL C. SUMMER NUCLEAR STATION INDIVIDUAL PLANT EXAMINATION OF l
EXTERNAL EVENTS (IPEEE) SUBMITTAL Fire events
- Post-FIVE Analysis"is 4.1E-4 per year. The fire CDF estimate remains high in comparison to other pressurized water reactor plant IPEEE submittals the NRC has reviewed to date.
However, you have identified no fire-related vulnerabilities, and implemented no plant improvements to reduce fire risk. You provided a qualitative dis:ussion of conservatisms in your analysis.
Please define plant " vulnerability" that you used in assessing the results of the IPEEE fire analysis. Describe how you determined that there are no cost-effective plant modifications that you could make to reduce fire risk at VCSNS.
- 2. Section 4.0 of your submittal provides " vulnerability evaluations" based on a qualitative discussion of the sources of conservatism in the unscreened fire areas / compartments analysis. The discussions included cited conservatisms which are derived primarily from three elements. One element was equipment duty factors. However, cuj factors are inherently included in the experience-based fire event data used in the study. A second element was giving some credit to suppression and detection. You based these credits on assessing damage times versus detection / suppression times. However, suppression has also been inherently credited through the severity factors used in the post-FIVE analysis (see related discussions in RAI items 7 and 8). The third element was the availability of attemate equipment to provide the backup system function. You should use appropriate plant models to support your assessment, if you take credit for such equipment within the context of a fire risk assessment. The models should include the credited equipment, appropriate human reliability analyses to ensure that required operator actions are appropriately accounted for, and cable tracing to ensure independence from the postulated fire.
Please provide a revised quantitative assessment of the CDF contribution taking into consideration the above three factors for each of the unscreened areas / compartments identified in Table 11 of yet submittal. Inc:ude a quantitative treatment of perceived major conservatisms using available fire risk analysis tools. Based on the results, please reassess the potential that these scenarios might represent fire vulnerabilities, and identify potential measures to reduce the fire risk at VCSNS.
- 3. The VCSNS Fire Emergency Procedures cited in the IPEEE study specify establishing power to the emergency bus from its respective diesel generator (A or B train) and tripping offsite power. In particular, page 52 of the submittal states that the VCSNS Fire Emergency Procedures (FEP 1.0, FEP-2.0, FEP-3.0, and FEP-4.0) direct the operators to isolate the offsite power source to the emergency bus (s) that is (are) supplying power to the designated safe shutdown train (s). After initiating safe shutdown, offsite power can be reestablished to selected emergency loads.
Please provide a copy of the above cited procedures (FEP 1.0 - 4.0). Assess the impact on the CDF values reported in Table 11 of not isolating offsite power from the emergency bus when required by the plant Fire Emergency Procedures.
- 4. - FEP 4.0 deals with fires requiring main control room (MCR) evacuation at VCSNS. The IPEEE discussion of MCR evacuation events (page 71 of the submittal) indicated that "it (FEP 4.0) assumes a loss of offsite power and no load sequencer actions." It is uncleai f
what this statement implies in the context of the IPEEE analysis. In particular, we need to know what was assumed in the IPEEE analysis regarding the plant state at the time of a forced MCR evacuathn. Operator actions are required to recover the emergency loads in the event of a " loss of offsite power with no load sequencer action." Failure to recover the emergency loads properly could lead to isolating both normal and emergency altemating current (AC) power for some period of time.
As a further note, page 54 of the submittal states that the load sequencer is located in the relay room (fire area CB-6). Accordingly, fires in this area have the potential to fail the load sequencer. This fire area is also identified on page 71 as one of the areas in which an
" uncontrolled fire" would force MCR evacuation. Thus, assuming loss of the sequencer function for such fires would be appropriate. For other fire areas / compartments, it is unclear whether or not loss of the sequencer has been, or should be, assumed in the IPEEE analysis.
For each area / compartment in which fires might lead to MCR evacuation at VCSNS, please provide (1) the assessment details with regard to the assumed state of the plant at the time of MCR evacuation, and (2) the operator actions required to shut down the plant from the control room evacuation panel. Also detail the human reliability analysis (HRA) for these scenarios including the operator action event trees, a description of the SAIC HRA methodology as applied to these scenarios, and the input values and their bases as applied to these HRA models. Provide an assessment of the impact on fire compartment CDF if you assume that all AC power is lost in the relevant MCR evacuadon scenarios, if the potential for sequencer failure (due to either fire-induced failures or random failures) has not been included.
- 5. Offsite power restoration failures was cred'ted in the
- Post-FIVE" analysis for fires in the MCR (pages 58 and 80). Table 12 of the VCSNS IPEEE submittalindicates that three panels (6116,6118, and 6225) associated with offsite power are located in the MCR. It is unclear whether or not the offsite power restoration failure prot, ability (OSPREC) shown in the event tree on page 7g of the submittal includes the potential that fire-induced failures in or.e (or more) of these panels might prevent restoring offsite power.
l
i Please provide a detailed assessment of the potential for restoring offsite power if any or all of the above listed panels are subject to fire damage. Also provide the basis for the assumed OSPREC value for MCR fire scenarios at VCSNS. Reassess the CDF contribution for the relevant fire scenarios in which fire damage might occur, if the potential for restoration might be impacted, and this potential was not included in the original analysis.
- 6. Licensees have misinterpreted and inappropriately extrapolated test results for the control cabinet heat release rate (HRR) in the Electric Power Resevch Institute (EPRI) Fire PRA implementation Guide.. The Guide uses cabinet HRRs as low as 65 Btu /sec. In contrast, experimental work has developed HRRS ranging from 23 w 1171 BTU /sec. Licensees should use an HRR in the mid-range of the currently avaliable experimontal data (e.g.,550 Btu /sec) for the analysis, considering the range of HRRS that could apply to different control cabinet fires, and to ensure that cabinet fire areas are not prematurely screened out of the analysis.
Please discuss the HRRs used in the VCSNS IPEEE assessment of control cabinet fires.
Discuss changes in the IPEEE fire assessment results, if you assumed that the HRR from a cabinet fire is increased to 550 Btu /sec.
- 7. The fires analysis in the VCSNS MCR appears to have included " double counting" in applying certain reduction factors. In particular, you applied a ? fire severity factor" of 0.05 to account for the fraction of fire events that result in a fully developed fire. At the same time, you applied a non-suppression probability of 3.4E-3 to account for the suppression of a panel fire prior to MCR evacu ation. These factors account for the same phenomena, namely, interruption of the fire prior to extended fire involvement. Hence, using one factor or the other might be considered appropriate with pmperfustification, but using both factors simultaneously in a fire risk analysis is generally inappropriate.
Further, the non-suppression probability of 3.4E-3 is based on the availability of optimally placed smoke detectors in each of the potential fire source panels. The IPEEE submittal described the smoke detectors present in the " main control board." However, there is no discussion of smoke detectors located in other panels. All MCR panels represent potential fire sources that might lead to MCR evacuation.
In response to these concems, please (1) select and justify applyireg the non-cuppression probability based on the physical configurations of each panel considered as a potential MCR fire source, and (2) reassess the fire CDF estimates for the MCR assuming that either the non-suppression or severity factor applies, but not both.
- 8. Page 51 of the IPEEE submittal indicates that you developed severity factors for various fire sources and scenarios in the Post-FIVE analysis. However, it appears from the discussions on page 67 that you applied both severity factors and automatic / manual non-suppression factors to at least some fire scenarios. As discussed in RAI number 7, applying both a severity factor as developed in the submittal and independent credit for other suppression efforts represents " double counting." It is also unclear for individual fire areas /compartr.1ents how the IPEEE has applied these factors in going from the FIVE results of Table 3 to the Post-FIVE results in Table 11.
e For each area / compartment considered in Table 11 of the IPEEE submittal, please provide a detailed breakdown of the contributing factors that comprise the estimated CDF. Identify the following:
original fire frequency from the FIVE analysis modified Post-FIVE fire frequency (Post-FIVE Step 1 results) e applied severity factor (s) (the Post-FIVE Step 2 results) credit taken for detection and suppression
=
. conditional core damage probability For those areas / compartments where the total fire-induced CDF is the sum of more than one individual scenario, provide the information for each contributing scenario.
- g. The automatic suppression system failure analysis used reliability values from the FlVE methodology. This data is acceptable for systems that have been designed, installed, and maintained in accordance with appropriate industry standards, such as those published by National Fire Protection Association (NFPA).
Please verify that automatic fire suppression systems at VCSNS meet applicable NFPA standards.
- 10. The heat loss factor (HLF) is defined as the fraction of energy released by a fire that is transferred to the enclosure boundaries. This is a key parameter for predicting component damage, as it determines the amount of heat available to the hot gas layer (HGL). A larger HLF means that a larger amount of heat (due to a more severe fire, a longer buming time, or both) is needed to cause a given temperature rise. If the value assumed for the HLF is unrealistically high, fire scenarios can be improperly screened out. Figure 1 provides a representative example of how HGL temperature predicticns can change assuming different HLFs. Please note that: (1) the curves are computed for a 1000 kW fire in a 10 meter x 5 meter x 4 meter compartment with a forced ventilation rate of 1130 cubic feet per meter; (2) the FlVE-recommended damage temperature for qualified cable is 700*F for qualified cable and 450'F for unqualified cable; and, (3) the Society for Fire Protection Engineers (SFPE) curve in the figure is generated from a correlation provided in the SFPE Handbook (Ref.1).
Based on evidence provided by an old American Society of Mechanical Engineers (ASME) research paper by Cooper et al. (Ref. 2), the EPRI Fire PRA implementation Guide recommends a HLF of 0.g4 for fires with durations greater than five minutes, and 0.85 for
" exposure fires away from a wall and quickly developing hot gas layers." However, as a general statement, this appears to be a misinterpretation of the research results. Cooper's paper (Ref. 2),' which documents the research results of multi-compartment fire experiments, indicates that the higher HLFs are associated with the movement of the HGL from the burning compartment to adjacent, cooler compartments. Earlier in the experiments, where the HGL is limited to the burning compartment, Cooper reports much lower HLFs (on the order of 0.51 to 0.74). These lower HLFs are more appropriate when analyzing a single compartment fire. In summary, (a) HGL predictions are very sensitive to the assumed value of the HLF; and (b) large HLFs cannot be justified for single-room scenarios based on the information referenced in the EPRI Fire PRA Implementation Guide.
I
I In the VCSNS submittal, you assumed the HLF to be 0.7 for 94% of the compartments.
However, the submittal states that, for 6% of the compartments, the HLF was relaxed to 0.85.
Please identify those fire areas where you used a HLF of 0.85 in the IPEEE analysis. For each of these fire areas, either: (1) Justify the value used and discuss its effect on identifying fire vulnerabilities, or (2) repeat the fire risk analysis using a more justifiable value and provide the resulting change in scenario contribution to CDF.
References for Question 10: 1) P.J. DiNenno, et al, eds., "SFPE Handbook of Fire Protection Engineering," 2nd Edition, National Fire Protection Association, p. 3-140,1995;
- 2) L. Y. Cooper, et al, "An Experimental Study of Upper Hot Layer Stratification in Full-Scale Multi room Fire Scenarios," ASME Joumal of Heat Transfer,1Qd,741-749, November 1982.
- 11. The IPEEE submittal discussed the existence of unprotected cable trays and conduits.
However, the submittal did not discuss applicable protection requirements and the electrical raceway fire barrier rating used in protecting cable trays and conduits.
Please (1) identify the types of electrical raceway fire barriers used at VCSNS; (2) identify all fire scenarios in which you take credit for electrical raceway fire barriers; and (3) assess the impact on estimated fire CDF if you do not take credit for these electrical raceway fire barriers.
- 12. Both fire-induced damage and automatic suppression system activation times for some of the critical fire scenarios in the VCSNS IPEEE submittal are so short as to be physically unrealistic. While it can be reasonably assumed that predicted short times to fire damage are conservative, the combined effect of short times for both critical damage and automatic suppression system activation might be overly optimistic. Section 2.5 of the IPEEE indicates that you assumed discharge delay times ranging from 10 to 30 seconds; this may be optimistic. For example, you assumed that pre action sprinklers have a 10-second delay time. However, systems compliant with NFPA 13 typically have discharge delays in the range of 1 to 3 minutes following detection system activation.
Please provide an assessment of the impact on the VCSNS fire CDF using more realistic automatic suppression actuation delay times. One option to achieve this would be to base the assumed timing value on the documented results of actual discharge time observed during the functional acceptance testing. A second acceptable approach would be to uniformly increase the actuation times to 3 minutes.
- 13. The EPRI Fire PRA implementation Guide assumes that all enclosed ignition sources cannot lead to fire propagation or other damage. The Guide also assumes that fire spread to adjacent cabinets cannot occur if the cabinets are separated by a double wall with an air gap, or if the cabinet in which the fire originates has an open top. Based on the discussion presented on pages 61 and 62 of the submittal, you applied these assumptions in the VCSNS IPEEE. This resulted in a number of fire sources being screened from the analysis.
These assumptions may be optimistic for certain types of electrical fires including oil-filled transformers, high-voltage cabinets, motor control centers, and switchgear. This is because i
an explosive breakdown of the electrical conductors may breach the integrity of the l
)
4 l
item / cabinet and allow fire to spread to combustibles located above the item / cabinet. For example, switchgear fires at Yankee-Rowe in 1984 and Oconee Unit 1 in 1989 both resulted in fire damage outside the cubicles.
Please provide the basis for assuming that all enclosed ignition sources cannot lead to fire propagation or other damage at VCSNS. Also discuss of how you aneiyzed each of the specific enclosures to conclude that the assumption applies to them.'
- 14. The description of the fire compartment interaction analysis (FCIA) on rge 13 of your submittaiindicates that you supplemented the FIVE barrier screening criteria with four additional screening criteria. These criteria are not consistent with the NRC-accepted FIVE methodology guidance.
ide,1tify each of the areas (if any) in Tables 2 and 3 of your IPEEE submittal where you used the additional criteria as the basis for dismissing the potential for fire spread between compartments in the FCIA. For each such set of areas, please provide a detailed assessment including the CDF contributions if these criteria are not applied and the.
impacted areas are combined for analysis.
- 15. It is important that the human error probabilities (HEPs) used in the analysis properly reflect the potential effects of fire (e.g., smoke, heat, loss of emergency lighting), even if these effects do not directly cause equipment damage in the scenarios being analyzed. The HEPs may be optimistic and result in the improper quantification or screening of scenarios if -
these effects are not treated. Page 28 of your IPEEE submittal indicates that you included '
HEPs in the system analysis, but implies that you took the values from the intemal events analysis (IPE). HEPs which are conservative with respect to an internal events analysis could be optimistic with respect to a fire risk analysis.
Please clarify whether you took the HEPs directly from the VCSNS IPE analysis, or modified them to reflect the unique aspects of each fire scenario at VCSNS. If the HEPs.
did not include fire effects (e.g., smoke, heat, loss of lighting) please provide an assessment of the impact on estimated fire-induced CDF if fire effects are included in the formulation of HEPs for each fire scenario.
Seismic events
- 1. As requested in the Generic Letter 88-20 (Supplement 4) and in NUREG-1407, Sections 6.2.2.1 and 6.3.3.1, please discuss the dispositioning of GSI-131, which deals with the seismic evaluation of the in-core flux mapping system for Westinghouse plants.
- 2. As requested in Appendix 4 of Generic Letter 88-20 (Supplement 4), please discuss seismic / flooding interaction issues, seismic actuation of fire suppression systems, and seismic degradation of fire suppression systems.
actions.. Regarding non-seismic failures and human actions, Section 3.2.5.8 of NUREG-l 1407 states that " success paths are chosen based on a screening criterion applied to non-seismic failures and needed human actions. It is important that the failure modes and human actions are clearly identified and have low enough probabilities to not affect the i
seismic margins evaluation." Please describe how you treated non-seismic failures and human actions in the VCSNS IPEEE, and address the concerns of the above quoted NUREG-1407 statement.
- 4. The turbine building and the steel roof structure on the auxiliary building that are Category ll structures were originally designed to the Uniform Building Code (UBC),1973 Edition.
Although you performed additional analyses to show their capacity to resist SSE seismic loading, the screening criteria based on Table 2-3 of EPRI NP-6041 for Category 11 structures require that Category ll structures meet the UBC Code,1985 Edition, Zone 4 requirements. Please verify to us that you meet the UBC Code,1985 Edition, Zone 4 requirements for these structures.
Hioh winds. floods. and other extemal events (HFOs) -
There are no RAls in this area.
I Figure 1 -
Sensitivit:' of hot gas layer temperature predictions to the assumed heat loss l
factor l
Time-Temperature Curves gog,
..r, il [
800 SFPi
-e-H.c = 0 70 700-
-*- M = 0 85
- m.
16:.- 600 -
-:p i ;&a
+-
+ H F = 0 94i y
e 4:'i
-x-H.F = 0 99 -"-"
5
- ~O 3 500 g
E 400 W-o
>=
300-
- .
- J
- r 200 100,, _,
g.v :x wx x.X XkX:XN* *'**
o, o
o o o o o o e o o o e e o ea o m o n w o e o m n a o c
=
ha m
9 D h M
o o r_e n 9
V_)
'Q R
Time (s) l e
Mr. Gary J. Taylor VIRGIL C. SUMMER NUCLEAR STATION South Carolina Electric & Gas Company cc:
l Mr. R. J. White J.
B.
Knotts. Jr.
Esquire Nuclear Coordinator Winston & Strawn Law Firm S.C. Public Service Authority 1400 L Street. N.W.
c/o Virgil C. Summer Nuclear Station Washington, D.C.
20005-3502 Post Office Box 88. Mail Code 802 Jenkinsville. South Carolina 29065 Resident Inspector / Summer NPS c/o U.S.~ Nuclear Regulatory' Commission Route 1. Box 64 Jenkinsville. South Carolina 29065 Regional Administrator Region II U.S. Nuclear Regulatory Commission Atlanta Federal Center 61 Forsyth Street. SW. Suite 23T85 Atlanta, Georgia 30303 Chairman. Fairfield County Council Drawer.60 Winnsboro. South Carolina 29180 Mr. Virgil R. Aut.'y Director of Radioactive Waste Management Bureau of Solid & Hazardous Waste Management Department of Health & Environmental Control 2600 Bull Street Columbia South Carolina 29201 Mr. Robert M. Fowlkes. Manager Operations South Carolina Electric & Gas Company Virgil C. Summer Nuclear Station. Mail Code 303 Post Office Box 88 Jenkinsville. South Carolina 29065 Ms. April R.' Rice, Manager Nuclear Licensing & Operating Experience l
South Carolina Electric & Gas Company l
Virgil C. Summer Nuclear Station. Mail Code 830 L
Post Office Box 88 l
Jenkinsville. South Carolina 29065