RS-17-032, Response to Request for Additional Information Regarding the License Amendment Request to Utilize the TORMIS Computer Code Methodology
ML17079A130 | |
Person / Time | |
---|---|
Site: | Byron ![]() |
Issue date: | 03/20/2017 |
From: | Gullott D Exelon Generation Co |
To: | Document Control Desk, Office of Nuclear Reactor Regulation |
References | |
RS-17-032 | |
Download: ML17079A130 (30) | |
Text
'1300 W111flclcl Ro:icl W:irrc11v1llc. IL 60555 Exelon Generation 630 657 2000 Off1cc RS-17-032 10 CFR 50 .90 March 20, 2017 U.S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington, DC 20555-0001 Byron Station, Units 1 and 2 Renewed Facility Operating License Nos. NPF-37 and NPF-66 NRC Docket Nos. STN 50-454 and STN 50-455
Subject:
Response to Request for Additional Information Regarding the License Amendment Request to Utilize the TORMIS Computer Code Methodology
References:
(1) Letter from D. M. Gullatt (Exelon Generation Company, LLC) to U. S. NRC, "License Amendment Request to Utilize the TORMIS Computer Code Methodology," dated October 7, 2016 (2) Email from J. S. Wiebe (U.S. NRC) to J. A. Bauer (Exelon Generation Company, LLC}, "Preliminary RAls for Byron Station TORMIS License Application Request," dated January 19, 2017 In Reference 1, Exelon Generation Company, LLC, (EGC) requested an amendment to Renewed Facility Operating License Nos. NPF-37 and NPF-66 for Byron Station, Units 1 and 2. This amendment request proposed to revise the Byron Station licensing basis for protection from tornado-generated missiles. Specifically, the Updated Final Safety Analysis Report (UFSAR) will be revised to identify the TORMIS Computer Code as the methodology used for assessing tornado-generated missile protection of unprotected plant structures, systems and components (SSCs) and to describe the results of the Byron Station site-specific tornado hazard analysis. The results from the Byron Station TORM IS analysis will be used to credit unprotected equipment for post-tornado safe shutdown. Of particular note, the Essential Service Water Cooling Tower (SXCT) fans and cells that survive a tornado strike will be credited for Ultimate Heat Sink (UHS) cooling as opposed to the original licensing basis that assumed all the unprotected SXCT fans are damaged by tornado-generated missiles.
In Reference 2, the NRC requested that EGC provide additional information to support their review of the subject License Amendment Request. During a February 2. 2017 follow-up clarification call to discuss details of the information request, it was agreed that EGC would provide the requested information to the NRC in 45 days; i.e., on or before March 20. 2017. The requested information is provided in Attachment 1 to this letter.
EGC has reviewed the information supporting the No Significant Hazards Consideration and the Environmental Consideration that was previously provided to the NRC in Reference 1. The
March 20, 2017 U. S. Nuclear Regulatory Commission Page 2 additional information provided in this submittal does not affect the conclusion that the proposed license amendment does not involve a significant hazards consideration. This additional information also does not affect the conclusion that neither an environmental impact statement nor an environmental assessment need be prepared in support of the proposed amendment.
In accordance with 10 CFR 50.91 , "Notice for public comment; State consultation," paragraph (b),
EGC is notifying the State of Illinois of this additional information by transmitting a copy of this letter and its attachment to the designated State Official.
As noted in Reference 1, EGC requests approval of this proposed license amendment by October 7, 2017.
This letter contains no new regulatory commitments. If you have any questions concerning this letter, please contact Joseph A. Bauer at (630) 657-2804.
I declare under penalty of perjury that the foregoing is true and correct. Executed on the 2o*h day of March 2017.
Respectfully, David M. Gullatt Manager - Licensing Exelon Generation Company, LLC : Response to Request for Additional Information -1: Draft Markup of Updated Final Safety Analysis Report Pages -2: Updated LAR Table 5 cc: NRC Regional Administrator, Region Ill NRC Senior Resident Inspector, Byron Station Illinois Emergency Management Agency - Division of Nuclear Safety
ATTACHMENT 1 Response to Request for Additional Information In Reference 1, Exelon Generation Company, LLC, (EGC) requested an amendment to Renewed Facility Operating License Nos. NPF-37 and NPF-66 for Byron Station, Units 1 and 2.
This amendment request proposed to revise the Byron Station licensing basis for protection from tornado-generated missiles. Specifically, the Updated Final Safety Analysis Report (UFSAR) will be revised to identify the TORMIS Computer Code as the methodology used for assessing tornado-generated missile protection of unprotected plant structures, systems and components (SSCs) and to describe the results of the Byron Station site-specific tornado hazard analysis. The results from the Byron Station TORMIS analysis will be used to credit unprotected equipment for post-tornado safe shutdown. Of particular note, the Essential Service Water Cooling Tower (SXCT) fans and cells that survive a tornado strike will be credited for Ultimate Heat Sink (UHS) cooling as opposed to the original licensing basis that assumed all the unprotected SXCT fans are damaged by tornado-generated missiles.
In Reference 2, the NRC requested that EGC provide additional information to support their review of the subject License Amendment Request. The requested information is provided below.
NRC Request for Additional Information In reviewing the Exelon Generation Company, LLC (Exe/on's) submittal dated October 7, 2016, related to use of the TORMIS methodology for the Byron Station (Byron), Unit Nos. 1 and 2, the NRG staff has determined that the following information is needed in order to complete its review:
RA/ 1 (DSSJ Tl1e licensee 's LAR references Regulatory Information Summary (RIS) 2008-14, "Use of the TORMIS Computer Code for Assessment of Tornado Missile Protection, " which includes reference to the 1983 TORMIS safety evaluation report (SER) (ADAMS Accession No. ML080870291). One of the five review points in the TORMIS SER describes that tornado characteristics sl10uld be employed for both broad regions and small areas around the site and most conservative values used. Based on tornado occurrence information in Byron Updated Final Safety Analysis Report (UFSAR) Section 2.3.1 .2.2, the probability of a tornado occurring within the one-degree square in which the Byron Station site is located is 21 .0 E-04 per year. The UFSAR indicates, " ... annual tornado probability for the Byron site area should be expressed as being in the range of .0010 to .0020, with a tornado return period of about 500 to 1000 years. "
The LAR specifies a total of 3, 289 tornadoes were reported in the 64 year period (i.e., 1950-2013) and calculates mean unadjusted occurrence rate of 3. 11 E-04 tornadoes I square mile I year. A correction for annual reporting trend is part of the TORMIS methodology and the adjusted occurrence rate to reflect the sub-region reporting trends is 3.58E-04 tornadoes I square mile I year. The LAR does not provide justification for the information above representing conservative values. nor does the UFSAR markup provided in the LAR indicate that prior information was updated to include new tornado frequency data that will represent new licensing basis information.
1of20
ATTACHMENT 1 Response to Request for Additional Information Provide a discussion on how t/1ese values were derived to represent the most conservative tornado frequency characteristics for both broad and small areas around the plant, including whet/1er t/1e values used are bounding and most conservative or justify accordingly. Also, provide a markup of the UFSAR to indicate changes in tornado frequency represented by the new methodology.
Response
As indicated in Section 3.4.5, "Compliance with the NRC Safety Evaluation Report (SER)
Acceptance Criteria," of the License Amendment Request (LAR) (Reference 1), a site-specific analysis was performed to generate a tornado hazard curve for Byron Station and a data set for the TORM IS analysis. The National Oceanic Atmospheric Administration (NOAA) Storm Prediction Center Severe Weather Database (i.e., SPC Database) was used as the basic source of data. This database includes corrections of the data over the years and consolidated event segments into single tornado events. The Byron Station TORMIS analysis is based on 64 years of NOAA data from 1950 to 2013.
To develop the tornado frequency characteristics for Byron Station, a broad 15° longitude x 15° latitude square, centered longitudinally at the plant, was used as the starting region. This large area covered 644, 752 square miles and included 11,804 tornados without a zero ending point in the NOAA Storm Prediction Center data set. Within this broad region, the tornado risk was quantified for subareas of 1° longitude by 1° latitude (i.e., 1° x 1° blocks) and 3° longitude by 3° latitude (i.e., 3° x 3° blocks). A statistical method, termed Cluster Analysis, was used to determine how the distinct subarea blocks grouped into similar clusters of tornado risk. These procedures were performed separately for both the 1° x 1° block and 3° x 3° block areas. The 1° x 1° block cluster results coupled with the 3° x 3° block cluster results were used to select the final Byron Generating Station (BGS) tornado sub-region.
The Byron Station tornado sub-region includes 47 1° blocks that surround the site and has broad statistical homogeneity for tornado risk analysis. The Byron Station sub-region includes cells of higher risk. Thus the modeling approach used to develop the Byron Station sub-region meets the intent of the TORM IS SER requirement to use data on tornado characteristics for both broad regions and small areas around the plant.
As indicated in the LAR the mean unadjusted occurrence rate for Byron Station was calculated as:
v = 1]/toA = 3.11 E-04 tornadoes I square mile I year Where:
11 = number of tornados reported within the Byron Station Sub-region = 3,289 tn = time period = 64 years A =area of the Byron Station sub-region= 165,059 square miles The occurrence rate is adjusted to correct for reporting trend and tornado classification error and random encounter errors, in accordance with the TORMIS methodology. The reporting trend adjustment was calculated to be 1.115. A second adjustment is made to account for reported tornadoes with zero latitude-longitude end points that were not included in the statistical analysis. Tornadoes with zero latitude-longitude end points are typically small events, average less than 0.03 square miles in path area, and aggregate to only 2. 7% of the total 2 of 20
ATTACHMENT 1 Response to Request for Additional Information reported path area in the Byron subregion. The method to account for zero end point tornadoes was to increase the occurrence rate by 3% to account for the small tornado strike risk for these events. This adjustment eliminated the need to simulate many small events in TORMIS with very small path areas and unknown path directions. The final adjusted occurrence rate for Byron Station was calculated as:
v = 3.11 E-04 x 1.115 x 1.03 = 3.58E-04 tornadoes I square mile I year It should be noted that the UFSAR Section 2.3.1.2.2, "Tornadoes and Severe Winds," estimated mean tornado probability values of 21.0 E-04 per year and 0.0010 to 0.0020 are values with units of "tornadoes per year" as opposed to the "tornadoes per year per square miles" used in the TORMIS analysis. Using the UFSAR "1-degree square" annual tornado frequency values (i.e., 1.0 tornado/year for the period 1953 - 1962; and 2.1 tornado/year for the period 1955 -
1967) and the UFSAR area for the 1-degree square of approximate 3470 mi 2
- the occurrence rate for tornadoes expressed in units of "tornadoes per year per square mile" are:
v =1.0 tornado/year+ 3470 mi22 =2.88E-04 tornadoes/year-mi 2 (1953 - 1962 data) v =2.1 tornado/year+ 3470 mi =6.05E-04 tornadoes/year-mi 2 (1955 - 1967 data)
The UFSAR values are averages with very large uncertainties resulting from a few years of data. In addition, the UFSAR occurrence rate may not include a distinction for very small events, such as the zero end point tornadoes in the current SPC Database. The TORM ISLAR occurrence rate uses significantly more data (1950-2013) over a much larger area, which reduces the uncertainty associated with determining the tornado occurrence rate. Additionally, a significant amount of the data used in the Byron Station TORMIS analysis is in the modern era (1994- current). The modern era includes more emphasis on reporting and verification of tornadoes which results in higher quality data. Thus, the differences in the UFSAR and TORMIS occurrence rates are reasonable given the significant differences in the periods of record, quality of data used and consideration of small area events. The TORMIS tornado occurrence rate for the Byron site was developed considering broad regions around the site and includes the contributions of small blocks with higher average risk.
Tornado occurrence rate (tornadoes per square mile per year) is only one input in the development of a tornado windspeed frequency model (i.e., the annual frequency of windspeed exceedance) for a site. Tornado path characteristics and areas, Fujita/Enhanced Fujita (FIEF) scale distribution, and windfield model/wind swath model (i.e., windspeed vs path area relationships) within the tornado path are some of the other key modeling inputs. A practical way to assess the conservatism of the Byron Station TORM IS tornado occurrence rate modeling is to compare the final windspeed frequency results to other independently developed windspeed frequencies for the site. This was done in LAR Attachment 1-2. Figure 8, "TORMIS Simulation of Byron Station Tornado Hazard for Plant Safety Envelope."
The Byron Station site-specific tornado risk used in the TORMIS analysis has been shown to be conservative when compared to the NRC Region I criteria given in NUREG/CR-4461 , "Tornado Climatology of the Contiguous United States, (PNNL-15112, Rev 2)," Revision 2. As shown in LAR Attachment 1-2, Figure 8, the Byron Station TORMIS tornado hazard exceedance probability values are conservative when compared to the NUREG/CR-4461 values.
3 of 20
ATTACHMENT 1 Response to Request for Additional Information The following table compares the exceedance probability values from NUREG/CR-4461 Table 8-1, "Recommended Tornado Design Wind Speeds," for Region 1 and the BGS TORRISK curve.
Peak Gust NUREG Table 8-1, BGS TORRISK BGS/NUREG Winds peed Exceedance Exceedance (mph) Probability, Region 1 Probability (from LAR Figure 8) 160 1.0 x 10*5 2.5 x 10*5 2.5 200 1.0 x 10*<> 3.4 x 10*<> 3.4 230 1.0 x 10-' 1.9 x 10-' 1.9 This comparison shows that the BGS wind speed exceedance probabilities have margin when compared to the NUREG/CR-4461 Region 1 values.
The NRC used the results of NUREG/CR-4461 in Regulatory Guide (RG) 1.76, "Design Basis Tornado for Nuclear Power Plants," Revision 1.
The TORMIS tornado hazard curve is conservative when compared to the data/guidance the NRC considers generally acceptable; i.e., NUREG/CR-4461, for defining the design-basis tornado for a nuclear power plant. Therefore, as required by the 1983 TORMIS SER, justification has been provided for the tornado frequency values and the associated tornado
- hazard curve used in the Byron Station TORMIS analysis.
UFSAR Section 2.3.1.2.2 will be revised to acknowledge the tornado parameters and tornado frequency values used in TORMIS (see Attachment 1-1, "Draft Markup of Updated Final Safety Analysis Report Pages"). Note that TORMIS only applies to a limited set of unprotected targets; therefore, the approach used for the Byron Station UFSAR markups is to retain the existing UFSAR information that still applies to the majority of the plant and specify that alternate information applies for the probabilistic tornado missile risk analysis.
RA/ 2(055)
UFSAR 2.3.1.2.2 and UFSAR 3.3.1.1 indicate the vertical velocity distribution and gust factors employed for the wind velocities are in accordance with ANSI A58.1-1972. A footnote to UFSAR Table 3.5-4 specifies vertical impact velocities are taken equal to 80% of the horizontal impact velocities. UFSAR 3.5.2 states, "The fans and motors are not protected from vertical or near vertical missiles. "
It is unclear whether vertical missiles were addressed in TORMIS analysis.
Provide discussion on how vertical missiles are addressed for unprotected components and update UFSAR to clearly define protection from vertical missiles.
Response
The approved TORM IS methodology explicitly considers all x, y, and z components of the missile velocity vector. The 3-D simulations integrate the equations of motion for each missile 4 of 20
ATTACHMENT 1 Response to Request for Additional Information and the resulting trajectories include a continuum of trajectory paths, including horizontal, vertical, and oblique trajectory paths. NP-2005 Volume 2,Section IV, describes missile motion and orientation models. The missile velocity vector at the instant of impact: therefore, inherently includes vertical and horizontal motions of the missile. As such, a continuum of missile velocity vector orientations at impact can occur, including horizontal, near- horizontal, oblique, near vertical, and vertical.
The TORMIS analysis is intended as a supplement to the Current Licensing Basis, instead of a full replacement. As such, the existing UFSAR Table 3.5-4, "Impact Velocities of Design-Basis Tornado-Generated Missiles," was modified to include "Design Basis" in the title to distinguish it from probability based TORMIS. The intent of this change was to clarify that the values listed remain valid for structures previously evaluated; for example the Auxiliary Building walls.
The proposed revision to UFSAR Section 3.5.1.4, "Missiles Generated By Natural Phenomena,"
submitted in Reference 1, states the following : "The characteristics of tornado-generated missiles considered in the probabilistic tornado missile risk analysis (TORMIS) described in Byron only Section 3.5.5 are found in Reference 15." Note that Reference 15 is the Byron Station TORMIS Design Analysis. This statement implicitly addresses missile velocity vector orientation by virtue of referencing the TORMIS Design Analysis. No further UFSAR revisions to specifically address vertical missiles are necessary.
RA/ 3(DSS)
LAR Section 3.4. 1 indicates, "The results from tl1e finite element analysis were then used to develop critical velocities for the other Byron Station TORMIS missiles. For these selected targets, damage is evaluated by comparing the missile velocity to the damage threshold velocity for the particular missile type and target group. If the missile velocity meets or exceeds the damage threshold velocity, it is scored as damage." UFSAR Table 3.5-4 contains specific values for the licensing basis impact velocities of tornado generated missiles.
Confirm that the missile velocities. used in the finite element analysis and TORMIS. bound the impact velocities in Table 3.5-4 of UFSAR.
Response
TOR MIS does not use the missile velocities listed in Table 3.5-4 of the UFSAR. Instead .
TORM IS simulates 3-D missile trajectories by integrating the equations of motion. Simulated missile trajectories consider the physical properties of each missile, including missile dimensions, weight and aerodynamic shape as they are released into a simulated tornado wind field. This approach produces simulated missile velocities at impact that include site specific missile sources and target characteristics.
In TORMIS, the damage threshold velocity can be specified for each target and each missile type. This approach was built into TORM IS to allow for target-specific damage calculations where analysis outside of TOR MIS could be used to determine the missile impact velocity thresholds that could produce failure of the target. The threshold damage velocities are input to TORMIS for each target-missile pair.
5 of 20
ATTACHMENT 1 Response to Request for Additional Information A finite element analysis was used to generate missile damage threshold velocities for the following targets:
- UHS riser pipes
- Diesel Auxiliary Feedwater Exhaust Pipe
- Diesel Auxiliary Feedwater Exhaust Cover
- PORV tail pipes There was no attempt to show that the subject unprotected targets were capable of withstanding UFSAR missile impact design velocities. The UFSAR missile velocities are retained as they are applicable for the Design Basis structures (i.e., non-TORM IS targets). The purpose of the finite element analysis is to determine the missile hit velocity for each TORMIS missile that results in unacceptable crimping damage of the four subject targets.
The calculated critical impact velocity for each missile-target pair was used to determine if a TORMIS missile hit results in unacceptable target damage. The threshold velocity for missiles causing damage to the DAFP exhaust pipes, DAFP exhaust cover plates, UHS riser pipes, and PORV tailpipes were input into TORMIS as 90% of the critical velocities calculated in the finite element analysis.
As an example, the UFSAR 6-inch pipe missile horizontal impact velocity used for tornado barrier design is 230 fps. For the PORV tail pipes, a finite element analysis determined the critical velocity for a 284 lb, 6-inch pipe missile is 172 fps. The TORMIS input value for the crimping velocity for this target was conservatively specified as 155 fps (90% of 172 fps). Any TORMIS 6-inch pipe missile hit on the PORV tailpipes is scored as a failure (i.e., unacceptable crimping damage) if the TORMIS missile velocity is greater than 155 fps.
RA/4 (DSS)
One of the five review points in the TORMIS SER indicates F-scale tornado classification should be used in order to obtain conservative results. In accordance with UFSAR Section 2.3.1.2.2,
=
the current licensing basis windspeed for the Byron Station are rotational velocity 290 mph and maximum translational velocity= 70 mph. This is consistent with the 1974 version of RG 1. 76 referenced in UFSAR Section 3.3.3.
The LAR submittal states a tornado hazard curve for Byron Station was developed and the EF-scale wind speeds were used in this analysis in accordance with NUREGICR-4461, Revision 2. The LAR further states the use of EF scale is consistent with the recently endorsed positions of NRG Regulatory Guide 1. 76 t/1at are based on NUREGICR-4461, Revision 2. LAR Table 3.4. 5-1 specifies wind speeds for EF5 tornado of 200-230 mph, which could potentially not bound the UFSAR values.
Provide details of windspeeds used in TORMIS and how they meet UFSAR licensing basis wind speeds or justify acceptability. If the intent is to update the licensing basis to incorporate the use of EF Scale and latest revision of RG 1. 76. then explain how all differences in the RG revisions are addressed (i.e .. missile spectrum, wind speed. faster auto, etc .. .) and update the UFSAR accordingly.
6 of 20
ATTACHMENT 1 Response to Request for Additional Information
Response
The Byron Station TORMIS analysis uses EF-Scale wind speed ranges. The EF-Scale wind speeds are input to TORM IS in the variable WINDG, which has the units of ft/sec. The following wind speeds were used in the Byron Station TORMIS analysis:
Local Tornado EF-Scale Wind WINDC Input Wind Intensity Speed (mph) Speeds (fps)
EFO 65-85 95.3-126.1 EF1 86-110 126.1-162.8 EF2 111-135 162.8-199.5 EF3 136-165 199.5-243.5 EF4 166-200 243.5-294.8 EF5 201-234 294.8-343.3 Revision 1 of RG 1.76 is based on Revision 2 of NUREG/GR-4461 which utilized the Enhanced Fujita (EF) scale. RG 1.76 Revision 1 states:
"Meteorological and topographic conditions, which vary significantly within the continental United States, influence the frequency of occurrence and intensity of tornadoes. The NRG staff has determined that the design-basis tornado wind speeds for new reactors should correspond to the exceedance frequency of 10*7 per year (calculated as a best estimate),
thus using the same exceedance frequency as the original version of this regulatory guide.
The results of the analysis indicated that a maximum wind speed of 103 meters per second (m/s) [230 miles per hour (mph)] is appropriate for tornadoes for the central portion of the United States;"
Note that 230 mph is consistent with the maximum EF scale wind velocity used in the Byron Station TORM IS analysis. Use of the current NRG accepted design basis wind speeds in the TORMIS probabilistic risk evaluation provides an adequate level of conservatism.
As indicated in the NRG Safety Evaluation (SE) for Fermi Unit 2, dated March 10, 2014 (Reference 3):
"The licensee stated that the original Enhanced Fujita scale wind speeds were utilized in the analysis. The NRG adopted the EF scale in the positions of Regulatory Guide 1.76, Revision 1, that are based on NUREG/GR-4461, Revision 2. The NRG staff concludes that the use of the EF scale is acceptable."
The use of the EF scale wind speeds is limited to evaluation of unprotected equipment using TORMIS. There is no intent to update the entire licensing basis to utilize the latest revision of RG 1.76. The new UFSAR Section 3.5.5, submitted with Reference 1, has been further revised to specifically note use of the EF scale as shown in UFSAR Section 3.5.5.3.a (see underlined text in Attachment 1-1).
7 of 20
ATTACHMENT 1 Response to Request for Additional Information RA/ 5 (DSS)
As indicated in LAR Section 3.4.6 (2a). the essential service water (SX) makeup pumps are located in the River Screen House, wl1icl1 is not protected against tornado missiles. For the case of a tornado impacting the river screen house, the non-safety related onsite deep well pumps are used to provide makeup water. The safety related power supply for t11e deep well pumps is described in UFSAR Section 9.2.5.2.3 and describe specific defense in depth credit for flooding only. T/Je capability for the deep well pumps to provide tornado event SX makeup to tl1e UHS is discussed in UFSAR Section 9.2.5.3.1, and operational assumptions for initiation of the deep water [well] pumps is described with other assumptions for the essential service water cooling tower(SXCT) in Section 9.2.5.3.5 (g). Further, UFSAR Section 3.5.5 appears to describe that conduit in the Auxiliary Building south wall and associated cable vaults that support operation of the deep well pumps are not protected, and are considered as part of the TORMIS analysis (on Table 3.5.17).
Confirm the piping, electrical and infrastructure of the non-safety onsite deep well pumps are adequately tornado protected to provide makeup water to function as defense in depth for SX makeup pumps. If supporting equipment or power for the deep well pumps are not protected from the effects of a tornado, describe how loss of these pumps are considered in TORMIS, and t/1e resultant probabilities for loss of the 11eat sink function.
Describe why UFSAR Section 9.2.5.2.3 only credits the deep well pumps as alternate makeup for flooding scenarios. but not tornado related scenarios. Additionally, describe how t/Je deep well pumps are credited as alternate makeup in UFSAR Section 9.2.5.3. 1 for the effects from tornados if supporting equipment is not protected (as described in Table 3.5.17). Make any appropriate changes to tl1e LAR and UFSAR. as warranted to correct information in the current UFSAR revision .
Response
Portions of the piping, electrical and infrastructure of the non-safety onsite deep well pumps do not currently have an acceptable level of tornado protection to provide makeup water to fully function as defense in depth for the SX makeup pumps. This lack of protection is one of the motivations for submitting the LAR.
As a result of walkdowns and document reviews, the following items relevant to the deep well pump infrastructure were identified as not completely protected from tornado generated missiles. A brief description of the tornado protection deficiency is provided. As extensive modifications would have been required to bring them into compliance, these targets are analyzed with TORMIS (these targets are listed in Table 1 of Reference 1).
- SXCT Switchgear Room Ventilation Louvers (Division 11 and 12 only) - Missile hits on the ventilation openings could impact the ventilation for the SXCT switchgear room. On a loss of room ventilation, the switchgear providing power to the deep well pumps could potentially over heat.
- Embedded conduits in the Auxiliary Building South Wall (Division 11 and 12 only) - The conduits do not have the UFSAR described minimum concrete thickness for protection from 8 of 20
ATTACHMENT 1 Response to Request for Additional Information tornado missiles. The conduits are covered by 18" of reinforced concrete. Power to the deep well pumps could be lost if the cables in the conduit are damaged by tornado missiles.
- Deep Well Pump Enclosures - The thickness of the enclosure concrete walls and roof are less than the UFSAR described minimum concrete thickness for protection from tornado missiles.
- Cable Vaults - Division 11 (Bus 131Z) and 12 (Bus 132Z) -The thickness of the concrete roof for the vaults are less than the UFSAR described minimum concrete thickness for protection from tornado missiles. Cables providing power to the deep well pumps are in the vaults.
- Non-ESF Switchgear Room Conduits - Missiles that can pass thru the Non-ESF Switchgear Room ventilation openings could hit conduits containing the power supply cables for the deep well pumps.
Regarding buried piping; the UFSAR has previously addressed the piping under Section 3.5.2, "Systems to be Protected," which states:
"All Category I buried pipes on Byron/Braidwood sites and the Category II (Non-Safety Related) Well Water (WW) piping from the onsite wells and pumps to the essential service water cooling towers at Byron have adequate soil cover for protection from tornado-generated missiles. These pipes are buried to depths greater than required minimum depth of 4 feet 1 inch, determined using Young's method."
The loss of the deep well pumps is considered in TORMIS using Target Group Number 33, Deepwell Enclosures and Associated Conduits. As indicated in Table 2 of Reference 1, the failure logic for Target Group 33 is the union of targets 60, 62, 123, 124, 125, 127, 130, 132, 134, and 143- 149. Thus the damage frequency for the group is the damage frequency for one or more of the unprotected targets associated with both deep well pumps. The deep well pump damage frequency is directly added to the Unit 1 and Unit 2 damage frequency values separately from the UHS Target Group frequency values.
Therefore, upon LAR approval, the aggregate of the piping , electrical and infrastructure of the non-safety onsite deep well pumps will have been shown to be adequately addressed regarding tornado protection.
UFSAR Section 9.2.5.2.3, "Category II Deep Well Pumps," is a brief summary of the Deep Well pumps and as such does not discuss tornadoes. The tornado impact on the deep well pumps is discussed under Section 9.2.5.3, "Safety Evaluation," Subsection 9.2.5.3.1, "Ultimate Heat Sink Design Basis." Similar to the above discussion, EGC recognized that the deep well pump infrastructure was not adequately protected from tornadoes and added them to the LAR (i.e.,
Reference 1) and UFSAR Table 3.5-17, "Targets Evaluated in TORMIS Analysis." Therefore, upon LAR approval, the aggregate of the piping , electrical and infrastructure of the non-safety onsite deep well pumps will have been shown to be adequately addressed regarding tornado protection and no additional UFSAR or LAR changes are required.
9 of 20
ATTACHMENT 1 Response to Request for Additional Information RA/ 6 (DSS)
RIS 2008-14 discusses t/1at the application should consider those cases where tornado missile damage to unprotected nonsafety-related SSCs that could adversely impact safety related SSCs. The LAR describes that the unprotected non-safety related Condensate Storage Tanks (CST) and piping from the CS Ts to the auxiliary feedwater (AF) pumps located in the Turbine Building are not included in the Byron TORMIS model. The safety related essential service water system is used as the backup suction source for the AF pumps if the CSTs or piping from the CSTs are damaged during a tornado event.
In the event of tornado, confirm failure of CST and secondary effects (e.g., flooding) will not adversely impact safety related SSCs.
Response
Two SSCs are sufficiently close to the CSTs to warrant consideration of secondary flooding effects; the OA deep well pump enclosure and the SXCT complex. The OA deep well pump enclosure is approximately 25 feet from the nearest CST. A walkdown of the subject area found that the ground slopes away from the CSTs and the deep well pump enclosure; this would provide preferential flow away from the enclosure. Given that the vulnerable feature of the enclosure (i.e., the vent opening) is over 6.5 feet above the focal ground level, it was concluded that the deep well pump enclosure would not be adversely impacted by flooding due to a tornado missile induced failure of the CST as the projected water level at the deep well pump enclosure would be far less than 6.5 feet above the local ground level.
The nearest point of the SXCT complex to the CSTs is 125 feet away. There is also a drainage ditch and a slightly raised roadway running the length of the SXCT complex. The combination of distance, drainage and obstructions support a conclusion that the SXCT complex would not be adversely impacted by flooding due to a tornado missile induced failure of the CST.
RA/ 7 (DSS)
The LAR markup of UFSAR Page 3. 5-23 contains a reduction in wall thickness for tornado barrier. The LAR UFSAR markup shows, "The walls and roofs of structures protecting the safety-related systems and components from design-basis tornado-generated missiles are of reinforced concrete with minimum thickness of "24" (changed to "2QJ and 14 inches respectively. The concrete used has a minimum cylinder strength of 3500 psi at 91 days." This change appears to be outside the scope of the TORM/S submittal.
Describe why this change is appropriate for consideration in this LAR.
Response
EGC agrees that the item is outside the scope of the TORMI S submittal. The item was included as a general markup for tornado related issues and is not related to TORM IS. The change will be processed separately under 10CFR50.59.
10 of 20
ATTACHMENT 1 Response to Request for Additional Information RA/ 8 (DSS)
The LAR states, "The peak 11eat input to the UHS for a post-tornado two unit slwtdown event (w/1ich assumes a dual unit loss of offsite power (LOOP)) is much less than tl1e peak heat load imposed on the UHS during a LOCA; therefore. fewer SXCT fans are needed for a post-tornado coo/down of bot/7 units; i.e., eit/1er 2 or 3 SXCT fans are needed depending on the case."
Section 9. 2. 5.3. 1. 1 of the UFSAR indicates, "The accident scenarios analyzed various single active failures and assumed that one or two essential service water cooling tower cells were initially out of service. "
Provide additional details on the analysis to verify only 2 SXCT fans are required to safely shutdown both units 1 & 2 during post-tornado conditions and to verify heat loading capability is adequate for post-tornado event shutdown.
Response
As discussed in LAR Section 1.0, "Summary Description," the LAR is requesting approval to revise the Byron Station licensing basis for protection from tornado-generated missiles. The TORM IS results will be used to credit un-protected equipment for post-tornado shutdown. After approval of the LAR, modifying the UHS licensing basis will be processed under 10 CFR 50.59.
Byron Station is not requesting NRC review and approval of the UHS temperature analysis; however, a summary of the UHS temperature calculation that will support the 50.59 evaluation is provided below.
As described in LAR Section 3.4.3, "Boolean Logic for the Ultimate Heat Sink." the number of cells that may be initially out of service is dependent on the outside air wet bulb temperature and number of operating units. The analysis assumes that two of the eight SXCT fans are initially out of service. An electrical random failure concurrent with the tornado event is assumed that results in the loss of power to two additional SXCT fans . Based on the results from the TORM IS analysis, two SXCT fans and four open risers without an operating fan survive the tornado and are considered to be available to provide cooling. The analysis determined that with the available fans and risers, and a wet bulb temperature less than or equal to 65°F, Cold Shutdown conditions can be reached for both units well before 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> (i.e .. within 36 hours4.166667e-4 days <br />0.01 hours <br />5.952381e-5 weeks <br />1.3698e-5 months <br />) while maintaining the SX supply water temperature less than or equal to the design maximum temperature of 100°F.
The analysis evaluates basin temperature vs. time by performing a minute by minute mass and energy balance on the SXCT basin.
As described in the LAR markup for UFSAR Section 3.5.4.1, "Missiles Generated by Natural Phenomena," the analysis uses SXCT performance curves generated for the 65° F wet bulb temperature using the method described in UFSAR Section 9.2.5.3.1.1 .2, "Steady State Tower Performance Analysis." This is the same method previously used for determining SXCT performance for a postulated loss of coolant accident.
As discussed in the LAR Section 3.4.3, "Boolean Logic for the Ultimate Heat Sink," the station will establish administrative controls to ensure that the assumed initial conditions in the post-11 of 20
ATTACHMENT 1 Response to Request for Additional Information tornado UHS cooldown analysis are met; i.e., the administrative controls will specify the number of SXCT fans required to be operable based on outside environmental conditions.
RA/ 9 (DSSJ Table 5 of LAR shows "Deepwel/ Enclosures" as containing Elect Rm 131Z and 132Z. but does not appear to include Elect Rm 231Z and 232Z. Additionally, Table 5 of LAR contains references to tables defined as "Source Table for Damage Frequency." The staff is unable to locate Table 2-3, 2-4, and 2-6 referenced in the table.
Clarify wl1ether this reference in Table 5 means, for example, "Table 2.6" or "Tables 2 through 6." If applicable, provide the missing Table references, or explain what is contained in the Tables sufficiently for the staff to evaluate the information in Table 5.
Response
The two deepwell pumps are powered from Unit 1 only. The OA pump is powered from Bus 131Z and the OB pump is powered from 132Z; therefore, electrical rooms 231Z and 232Z are not applicable to the deep well pumps.
Table 5 of the LAR was extracted directly from the calculation done by ARA. The source tables referenced are tables in the calculation. The cross-reference is as follows :
Table 2-3 = LAR Table 1 Table 2-4 = LAR Table 2
=
Table 2-6 LAR Table 4 LAR Table 5 has been updated to refer to the internal LAR table numbering scheme and is provided in Attachment 1-2 to this letter.
RA/ 10 (DSSJ Section 3.5.4.4 of the UFSAR markup in the LAR appears to make changes to Braidwood specific SSCs. Tl1e LAR is specific to changes proposed to Byron.
Justify what appears to be changes to Braidwood in Section 3. 5. 4. 4, or correct the markup to change Byron only UFSAR elements.
Response
The change from stacks or vent pipes to tailpipes for both Byron Station and Braidwood Station is outside the scope of the LAR and will be addressed separately as an editorial change.
Previously, Section 3.5.4.5 (now renumbered as Section 3.5.4.4) was applicable to both Byron Station and Braidwood Station. As Braidwood Station has not implemented TORMIS, the section was changed to "Braidwood only." A sentence was added to the end of the new Section 3.5.4.4 to refer to the new Section 3.5.5 (applicable to Byron Station only).
12 of 20
ATTACHMENT 1 Response to Request for Additional Information Note that, after further evaluation, it was concluded that all of the revisions made to the new Section 3.5.4.4 are outside the scope of the TORMIS LAR and will be addressed under the 50.59 process.
RA/ 11 (DRAJ RIS 2008-14 describes identified items that licensees should address when performing an approved TORMIS methodology per t/1e November 29, 1983, safety evaluation report (SER)
(Reference 1 in the Byron LAR). T/1e SER found that the methodology contained in EPRI NP-2005 (Reference 5 in the Byron LAR), is an acceptable approach for demonstrating compliance with the requirements of General Design Criteria (GOG) 2. In Section 2.0 of the LAR, the licensee cites that that the proposed revision to the Byron licensing basis is based on the NRG approved methodology in EPRI NP-2005, as well as other EPRI documents (NP-768 and NP-769).
In NP-2005.Section II. "Probabilistic Models and Simulation Methodology," TORMIS model treatment of a multiple reactor plant is described, specifically how to combine probabilities for a multiple reactor plant in detemJining the resultant damage probability. Table [Figure] 11-4 and Equations 34 and 35 are used to address multiple reactor targets.
Section 3. 4. 5 of the LAR describes how the licensee complies with the SER criteria. Item #5 of that section discusses any deviations from the TORMIS calculational approach. There is no discussion or justification referenced for difference in treating multiple unit plants as a different calculational probability, as described in the LAR Attachment 1 Tables 3.4.4-1and3.4.4-2, and Section 4.3, where damage frequency is shown as less than 1. 0 E-6 per year. per unit. Additionally, a calculated composite site damage frequency that exceeds 1.0 E-6 is shown on Tables 3.4.4-1and3.4.4-2. A precedent is cited with a D.C. Cook safety evaluation (Reference 12 in LAR). where a per reactor/per year result for a two unit plant was described as acceptable; however, the combination of both units probabilities remained below 1.0 E-6 for the TORMIS results, as noted in the staff's safety evaluation Justify how your calculation for a per unit/per year aligns with NP-2005. and whether this is a deviation from the model methodology. Additionally, provide infomJation for calculation of the composite site damage frequency, whether that aligns with NP-2005 calculations. and why that result should not be considered for meeting the probability threshold of 1.O E-6.
Response
The Byron TORMIS calculation damage frequency uses the damage frequency straight from the TORMIS methodology on a per year basis. NP-2005 Volume I, Figure 11-4, "Risk Aggregation for Multiple Time Periods," (found on page 11-30) and Equations 34-36 (found on page 11-33) address calculating the mean damage probability for multiple time periods such as different phases of plant construction for a multi-unit site. Specifically, Figure 11-4 illustrates the activities and operational sequence that might be associated with a 3-unit plant during construction/operation: and Equations 35 and 36 are used to estimate the mean per year damage probability over the entire time period. This discussion does not identify acceptance criteria nor does it discuss application of acceptance criteria on either a unit-specific or total site basis. The "Multiple Time Periods" illustrated by Equations 34-36 are computations done 13 of 20
ATTACHMENT 1 Response to Request for Additional Information outside the code using the per year results produced by TORMIS. The Byron Station TORMIS calculations for "multiple time periods" follow the TORM IS methodology.
Multiple time periods for Byron Station correspond to outage and non-outage conditions. Non-outage time periods have additional missile sources, which are treated by introducing additional potential missiles in the appropriate zones that reflect materials, equipment, new trailers, etc.
that exist during outage conditions. These potential outage missile sources are modeled in distinct TORMIS runs representing outage time periods. Three time periods were simulated with TORMIS : (1) Unit 1 in an outage state with Unit 2 operational; (2) Unit 2 in an outage state with Unit 1 operational ; and (3) Unit 1 and Unit 2 operational. Based on review of outage times, each unit was conservatively assumed to be in an outage condition 8.25% of the time. These three time periods were combined to calculate a per-unit and composite site damage frequency, consistent with the methodology in NP-2005. Thus, there is no deviation from the TORM IS methodology for computing per year damage frequency for individual units and/or the plant.
14 of 20
ATTACHMENT 1 Response to Request for Additional Information The damage frequency for each unit is calculated by summing the damage frequency for the unit specific target group with the damage frequency of common target groups. Thus the Unit 1 damage frequency in LAR Table 5 is calculated as the sum of the following target groups (common target groups noted with*):
Unit 1 Damage Frequency (Data from ll.eferencc I. Table 5)
Dm11a;.:e Freq11e11cy (.rr*')
In* Cll!it!
Target lde11t(fietl by BGS Corre.\po11tli111: Tllr1:et Group.'i I Cell Out 2 Cell.'i Out of'Seri*ice of'Seri*ice Li i IS Riser Pipes UHS Anti Vortex Roxes & Trash Screens UHS Fan Motor and Power Feeds Ul-IS Fan Gear Box Oil Gauges UHS Personnel Hatches Ultimate Heat Sink ( UHS)* 1.7J E-07 I .4'.! E-07 UHS Inspection Hatches UHS Fan Blades UHS Electrical Room Louvers Embedded Conduits in South Wall of Auxiliary Building Underground Cable Vaults Deepwell Enclosures, Electrical Rooms Deepwcll Enclosures 9.64E-08 131 Zand I 32Z, and related condu its*
Div 11 and 21 Conduit behind Conduits in Non-ESF Room I .60 E-09 Open in!!*
Arithmetic Sum of Common Target Groups 2.71E-07 2..&0E-07 Diesel Aux FW Pump Exhaust UI DAFP 3.16E-07 UI PORV 1 6.61E-IO UI PORV2 2.14E-IO PORV Tailpipes UI PORV 3 8.93E-I 0 UI PORV4 3.06E-IO UI MSSV I l.48E-09 Ul MSSV 2 3.07E-10 MSSV Tailpipes U 1 1vlSSV 3 5.79E-09 UI MSSV 4 7.59£-10 RWST Hatch U I RWST Hatch 2.45E-07 U 1 MEER Div 2 Opening 5.11 E-09 U I New Targets U I MEER Div I Opening 6.53E-08 UI CR HVAC Intake Opening 4.00E-10 Arithmetic Sum of Unit I Specific Target Croups (1.42E-07 6.42E-07 Arithmetic Sum of Unit I Target Groups+ Common Tar2et Groups 9.IJE-07 8.83E-07 15 of 20
ATTACHMENT 1 Response to Request for Additional Information Similarly the Unit 2 damage frequency is the sum of the following Unit 2 target groups and common target groups:
Unit 2 Danrngc Frc<1ncncy (Dah1 from H.crcrcncc I, Table 5)
D1111111ge Freq11e119* (.rr-')
b~* Ctt.\*e Target ltle11t(/ietl l~I' IJGS Corre.'ipomlillg T11rget Grottps I Cell 0111 2 Cell.\* 0111 o/'Sen*ice o/'Serl'ice UI IS lfoer Pipes UllS Anti Vortex Ooxes & Trash Screens Ul-IS Fan Motor and Power Feeds UHS Fan Gear Box Oil Gauges Ul-IS Personnel Hatches Ultimate Heal Sink (Ul-IS)* 1.73E-07 I.42E-07 Ul-IS Inspection I-latches Lii IS ran Blades UHS Electrical Room Louvers Embedded Conduits in South Wall of Auxiliary Building nderground Cable Vaults Deepwell Enclosures. Electrical Rooms Deepwell Enclosures 9.64E-08 131 Z and I32Z. and related condu its*
Div 11 and 2 1 Conduit behind Conduits in Non-ESF Room l.60E-09 Open ing*
Arithmetic Sum of Common Target Grou1>s 2.71 E-07 2.40E-07 Diesel Aux FW Pump Exhaust U2 DAFP 3. 77E-07 U2 PORV I l.46E-10 U2 PORV 2 4.57E-IO PORV Tailpipes U2 PORV 3 4.29E-10 U2 PORV 4 4.75E-IO U2 MSSV I 5.63E-IO U2 MSSV 2 I. ISE-09 MSSV Tailpipes U2 MSSV 3 7.76E-IO U2 MSSV 4 I .36E-09 RWST Hatch U2 RWST Hatch 2.39E-07 U2 CR HV AC Intake Opening I.50E-09 U2 New Targets U2 l'v1EER Div I Opening 6.16E-08 U2 MEER Div 2 Opening I. I 2E-08 Arithmetic Sum of Unit 2 Specific Tar2et Grouos 6.96E-07 6.96E-07
.\rithmetic Sum of Unit 2 Tan?et Grou1>s +Common Target Groups 9.66E-07 9.36E-07 16 of 20
ATTACHMENT 1 Response to Request for Additional Information The composite site damage frequency is the sum of all of the Unit 1 target group values, Unit 2 target group values. and common target group values, as shown in the table below.
I Cell 2 Cell.\'
0111 o{Sen*ice Out of Serl'ice Arithmetic Sum of Common Tan!et Grou11s 2.71 E-07 '.!AOE-07
.\rithmetic Sum of ll nit I Specific T11rget Grou11s 6.42L-07 6.4'.! E-07
,\rithmelic Sum of Unit 2 Suecilic T:tr1?.cl Grou11s 6.% E-07 6.%E-07 Composite Site Dnmagc Frequency 1.61 E-06 I .58 E-06 Comparing the result of the Unit damage frequencies (as opposed to the composite site damage frequency) to the acceptance criteria probability threshold of 1.0 E-6 is appropriate based on the following rationale.
The TORMIS Safety Evaluation Report, dated October 26, 1983, and the approved EPRI methodology did not establish TORMIS acceptance criteria. The acceptance criterion was specified in an NRC memorandum from H. R. Denton (Director, Office of NRR) to Victor Stello (Deputy Executive Director for Regional Operations and Generic Requirements}, "Position on Use of Probabilistic Risk Assessment in Tornado Missile Protection Licensing Actions," dated November 7, 1983. This memorandum is cited in RIS 2008-14. These documents state:
" ... the guidance of SRP Section 2.2.3 is applicable to tornado missiles. This guidance, which we will use in our probabilistic tornado missile reviews, states that an expected rate of occurrence of potential exposures in excess of the 10 CFR 100 guidelines of approximately 1o-s per year is acceptable if, when combined with reasonable qualitative arguments, the risk can be expected to be lower."
As indicated in 10 CFR 100.11, "Determination of exclusion area, low population zone, and population center distance," paragraph (b}, for sites with multiple reactors, fission product releases may be applied on a per unit basis when the reactors are independent to the extent that an accident in one reactor will not initiate an accident in another. The independence of the Byron Station Units is documented in UFSAR Section 3.1.2.1.5, "Evaluation Against Criterion 5
- Sharing of Structures, Systems, and Components." Consistent with this UFSAR section, the TORM IS targets are "unit independent" to the extent that the unit specific TORM IS targets are not required for safe shutdown of the opposite unit (for example loss of the Unit 1 DAFP will not challenge safe shutdown of Unit 2) . With this level of independence, damage to a unit specific target on one unit does not increase the rate of occurrence of potential exposure of the other unit. W ith the common target group damage frequency added to each unit, the risk of release for each unit and the site is adequately quantified and it is appropriate to consider the acceptance criteria on a per unit basis.
Additional precedent can be found in the TORM IS SE for the Joseph M. Farley Nuclear Plant, dated September 26, 2001 (Reference 4). In the SE, the NRC agreed with the licensee's determination that the acceptance criteria may be applied on a per unit basis due to the independence of unprotected SSCs between the two units following a tornado event.
Application of the SRP guidance to the TORM IS results for damage frequency also contains inherent conservatism. This conservatism stems from the assumption that a tornado missile strike that results in "target damage" also causes a radioactive release rather than performing specific evaluations to determine whether the damage can actually cause a release.
17 of 20
ATTACHMENT 1 Response to Request for Additional Information RA/ 12 (DRA)
RIS 2008-14 describes identified items that licensees should address wl1e11 performing an approved TORMIS methodology per the November 29, 1983, safety evaluation report (SER)
(Reference 1 in the Byron LAR). Section 3.4.5 of the LAR describes l10w the licensee complies with tl1e SER criteria.
In Section 4.4, Conclusions, the licensee states "There are many additional aspects of the TORMIS modeling and inputs that ensure bounding and conservative results."
Clarify tl1e meaning of this statement and relevance to additional information that is appropriate for staff review. Provide a list of the additional modeling and input parameters on this TOR MIS analysis if not already included in the LAR. and evaluate these parameters per criteria listed in Section 3.4. 5.
Response
The intent of this statement was to generally summarize that various aspects of the modeling and input parameters already mentioned in the LAR ensure bounding and conservative results. The following are specific conservatisms noted in the LAR:
- Section 3.4.1 - In general, the finite element analysis missile models were built to be conservatively strong and rigid. Conversely, target models were built to be conservatively weak to maximize the degree of potential crimping.
- Section 3.4.1 - The threshold velocity for missiles causing damage to the DAFP exhaust pipes, DAFP exhaust cover plates, UHS riser pipes, and PORV tailpipes were input into TORMIS as 90% of the critical velocities calculated in the finite element analysis.
- Section 3.4.3 - For the evaluation of the UHS, the random electrical bus failure was assumed to affect the in-service cells. This assumption is conservative since the out of service cells could be affected by the random electrical failure vice the in-service cells.
- Section 3.4.3 - For the evaluation of the UHS, the cells with the lowest probability of tornado missile damage are the cells that are either out of service or have failed from the random electrical failure. This maximizes the calculated damage frequency for the UHS.
- Section 3.4.3 - The analysis conservatively assumes that a failure of any of the eight SXCT riser pipes results in the failure of all cells in the UHS.
- Section 3.4.6 - The Byron Station TORMIS analysis conservatively did not increase the size of missile shielding targets for offset hits. This approach produces a conservative result in compliance with the RIS comment on "tumbling missiles."
- Section 3.4.6 - The developed tornado hazard curve for Byron Station is conservative when compared to NRC Region I criteria given in NUREG/CR-4461, Revision 2.
As part of the LAR development, all of these items were previously considered with regard to compliance with the NRC TORMIS SER acceptance criteria.
18 of 20
ATTACHMENT 1 Response to Request for Additional Information Additionally, as discussed in the response to RAI Question 11. application of the SRP guidance to the TORMIS results for damage frequency also contains inherent conservatism. This conservatism stems from the assumption that a tornado missile strike that results in "target damage" also causes a radioactive release, rather than performing specific evaluations to determine whether the damage can actually cause a release.
RA/ 13 (DRA)
RIS 2008-14 describes identified items that licensees should address when performing an approved TORMIS methodology per the November 29. 1983, safety evaluation report (SER)
(Reference 1 in the Byron LAR).
In Section 3.4.2, "Boolean Logic Approach," the licensee states that hit and damage frequencies for groups of targets evaluated in TORMIS are commonly combined using Boolean operators ( U and n) to aid in summarizing the results and understanding the effects of the system redundancies. The union ( U) operator means that if any one of the targets is damaged in a tornado, the system is assumed to fail. The intersection (n) operator means that all the intersected components must be damaged in a tornado strike for the system to fail.
Combinations of union and intersection operators can be put together to describe multi-component system failure logic for plant systems and subsystems.
Clarify if tl1e calculation of any mean (cumulative) tornado missile damage probability uses any intersection (n) operator that requires damaging multiple targets simultaneously for establishing a damaged state. If multiple targets need to be simultaneously struck, please summarize the guidelines used to identify such groups and explain how they are modeled in TOR MIS.
Response
The intersection operator, that requires damaging multiple targets simultaneously for establishing a damage state, was used to calculate the TORM IS tornado missile damage frequency for the UHS target group. As described in LAR Section 3.4.3, success for the UHS is defined as 3 of 8 SXCT cells surviving for the "one cell out of service" case: or 2 of 8 cells surviving for the "2 cells out of service" case. With a postulated electrical bus failure causing failure of two cells, success is defined as 3, 4, or 5 of the remaining 5 cells surviving for the 1 cell out of service case and 2, 3, or 4 of the remaining 4 cells surviving for the 2 cell out of service case. Thus multiple targets need to be damaged during the simulated tornado to resu lt in failure of the UHS.
A listing of failure events affecting the survival of the U HS are defined in Attachment 1-1, Table 4 of the LAR. As an example; combination number 10 is Fan Hand Fan D damaged by tornado missiles. For the 1 cell out of service case, this results in 1 cell out of service+ 2 cells lost to the single failure+ 2 cells lost to tornado missiles with 3 cells surviving, which is not considered a failure. An additional UHS target would need to be damaged to result in a UHS target group failure.
As discussed in LAR Section 3.4.3. this logic was modeled in the analysis with the TORMIS post-processor TORSCR using the Boolean intersection (n) operator. As discussed in LAR 19 of 20
ATTACHMENT 1 Response to Request for Additional Information Section 3.4.2, TORSCR is a FORTRAN computer code that is used to post-process TORMIS output files. Its primary function is to compute Boolean combinations of target hit and damage probabilities over multiple targets.
Boolean logic may be applied to any group of targets that support a common function ; however, the intersection operator was not used for any targets other than the UHS in the Byron Station TORMIS analysis.
REFERENCES
- 1. Letter from D. M. Gullott (Exelon Generation Company, LLC) to U. S. NRC, "License Amendment Request to Utilize the TORMIS Computer Code Methodology," dated October 7, 2016
- 2. Email from J. S. Wiebe (U . S. NRC) to J. A. Bauer (Exelon Generation Company, LLC),
"Preliminary RAls for Byron Station TORMIS License Application Request," dated January 19, 2017
- 3. Letter from T. J. Wengert (NRC) to J. H. Plona (DTE Electric Company), "Fermi 2 -
Issuance of Amendment Re: Revise the Fermi 2 Licensing Basis Concerning Protection from Tornado-Generated Missiles (TAC NO. MF0497)," dated March 10, 2014
- 4. Letter from Frank Rinaldi (NRC) to D. N. Morey (Southern Nuclear Operating Company, Inc.), "Joseph M. Farley Nuclear Plant, Units 1 and 2 Re: Issuance of Amendments (TAC NOS. MA9495 AND MA9496)," dated September 26, 2001 20 of 20
ATTACHMENT 1-1 Draft Markup of Updated Final Safety Analysis Report Pages BYRON STATION UNITS 1AND2 Docket Nos. 50-454 and 50-455 Renewed Facility Operating License Nos. NPF-37 and NPF-66 REVISED UFSAR PAGES 2.3-6 (included for context) 2.3-7 (included for context) 2.3-8 2.3-52a 3.5-26c
BYRON-UFSAR formula for the total attractive area as given here assumes a lightning strike current intensity of 2 x 10 4 amperes with a 50 % frequency of occurrence.
For the Byron Station, the smallest rectangle enclosing the reactor containment buildings is approximately 132.3 meters in length and 45.7 meters in width (see Byron Drawings M-5 and M-14). The height of the containment building is approximately 60.7 meters. It has been assumed that the height of the entire rectangle is 60.7 meters. This issues a realistic estimate of a lightning strike on the containment structures. The attractive area of the rectangle surrounding the containment buildings is therefore approximately 0.095 km-.
The reactor containment buildings of Byron Station have a probability of being struck which is equivalent to:
- 8. 4 flashes x 0 . 095 km 2 = 0 . 798 flashes (2.3-2b) km 2 yr yr Hence, a conservative estimate of the recurrence interval for a lightning strike on the reactor containment buildings is:
1
= 1.25 years/flash (2.3-2c)
- 0. 7 98 flashes I yr The area of the Byron Station site is approximately 1000 acres, or about 4.0 km 2
- Hence the expected frequency of lightning flashes at the site per year is 34 as calculated below:
- 8. 4 flashes km2 = 34 flashes x 4.0 (2.3-3) km x year year 2.3.1.2.2 Tornadoes and Severe Winds Illinois ranks eighth in the United States in average annual number of tornadoes (Reference 11). Tornadoes occur with the greatest frequency in Illinois during the months of March through June. For the period 1916-1969, the publication "Illinois Tornadoes" (Reference 11) lists 38 tornadoes which occurred in the 8 county area (Ogle, Carroll, Stephenson, Winnebago, Boone, DeKalb, Lee, and Whiteside) surrounding and including the Byron Station site. Figure 2.3-1 shows the county distribution of tornadoes for the entire state for the same period of record. For Ogle County, the total number of tornadoes was six.
Tornadoes can occur at any hour of the day but are more common during the afternoon and evening hours. About 50% of Illinois tornadoes travel from the southwest to northeast. Slightly over 80% exhibit directions of movement toward the northeast through east. Fewer than 2% move from a direction with some easterly component (Reference 11).
2.3-6 REVISION 9 - DECEMBER 2002
BYRON-UFSAR The likelihood of a given point being struck by a tornado in any given year can be calculated using a method developed by H.
C. S. Thorn (Reference 12). Thorn presents a map of the continental United States showing the mean annual frequency of occurrence of tornadoes for each 1-degree square (latitude x longitude) for the period 1953-1962. For the 1-degree square containing the Byron Station site, (approximately 34 70 mi 2 in area), Thorn computed an annual average occurrence of 1.0 tornadoes. Assuming 2.82 rni 2 is the average area covered by a tornado (Reference 12), the mean probability of a tornado occurring at any point within the 1-degree square containing the Byron site in any given year is calculated to be .0008. This converts to a mean recurrence interval of 1230 years. Using the same annual frequency but an average area of tornado coverage of 3.5 rni 2 (from Wilson and Changnon, Reference 11), the mean probability of a tornado occurrence is .0010.
More recent data (Reference 8) containing tornado frequencies for the period 1955-1967 indicates an annual tornado frequency of 2.1 for the 1-degree square containing the Byron site. This frequency, with Wilson and Changnon's average path area of 3.5 rni 2 , results in an estimated mean tornado probability of .0021, with a corresponding mean return period of about 470 years.
The results were presented in order to provide a reasonable estimate of tornado probability without addressing the accuracy of the estimate. Because of uncertainties in regard to tornado frequency and path area data, the annual tornado probability for the Byron site area should be expressed as being in the range of .0010 to .0020, with a tornado return period of about 500 to 1000 years. However, a conservatively high estimate can be taken to be .0021 or 470 years.
For the period 1970-1977, the NOAA publication "Storm Data" lists 17 tornadoes which have occurred in the 8 county area (Ogle, Carroll, Stephenson, Winnebago, Boone, DeKalb, Lee, and Whiteside) surrounding and including the Byron Station site.
The majority of these tornadoes were short in length, narrow in width, and weak in intensity. Four tornadoes, however, were severe enough to cause damage losses in excess of $50,000 in each case.
The most destructive tornado recorded during the period 1970-1977 in the vicinity of the Byron Station occurred on April 6, 1972, near Polo in Ogle County. An estimated $200,000 in damage losses was left in the wake of the tornado.
Approximately 20 rural farm buildings, a home, and a trailer were severely damaged and four electrical transmission towers were bent to the ground. Winds to 100 mph were observed at Polo Airport where a hangar and a cement block building were destroyed. The structural damage implies that the maximum wind speed of this tornado was approximately 150 mph.
2.3-7 REVISION 9 - DECEMBER 2002
BYRON-UFSAR The tornado path length extended 40 km with an average width 45 me t ers. A conservative estimate of the t or n a do p a th a r e a i .
- 1. 8 km-.
The foll o wing are the design-basis tornado parameters (Referen c e 13) that were used for the Byron Station:
- a. rotational velocity = 290 mph,
- b. maximum translational velocity = 70 mph,
~ . radius of maximum rotational velocity = 1 5 0 f e et,
- d. pressure drop = 3.0 psi, and
- e. rate of pressure drop= 2.0 psi/sec.
The des i gn wind velocity used for Seismic Category I s tructu res a t the Byron Station site is 85 mph considering a 100-ye ar r ecurrence interval. For Seismic Category II structu re s, the g overning design wind velocity used is 75 mph with a recurr e nce i nterval of 5 0 years. The design wind velocities f or the 5 0-year and 100-year recurrence intervals are obtained from Figures 1 and 2 of the American National s tandard Building Cod e Requirements for Minimum Design Loads in Buildings" (Re ference 1 4). The vertical velocity distribution and gust fa cto r s e mployed for the wind velocities are from Reference 13 f or e xposure Type C (see Subsection 3.3.1.).
Th~ tornado pa rameters and tornado fre quency values used in the orobabilisti c tornado missile risk ana lysis (TORMIS) described in Byron only Sec tion 3.5.5 are found in Reference 41 .
2.3.1.2.3 Heavy Snow and Severe Glaze Storms Severe winter storms, those that produce snowfall in excess of 6 inches and often are accompanied by damaging glaze, are responsible for more damage in Illinois than any other form of severe weather, including hail, tornadoes, or lightning (Reference 15). These storms occur on an average of five times per year in the state. The state probability for one or more severe winter storms in a year is virtually 100% while the probability for three or more in a year is 87%. During the 61-year period-of-record 1900 to 1960 used in a severe winter storm analysis (Reference 15), a typical storm had an average point duration of 14.2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />. Data on the average areal extent of severe winter storms in Illinois~show that they deposit at least 1 inch of snow over 32,305 mi-, with more than 6 inches covering 7500 mi . The northwestern area of Illinois (including the Byron Station site) had 144 occurrences of a 6-inch snowstorm during the years 1 900-1960. About 60 of these storms deposited more than 6 inches of snow in the Ogle County area. These frequencies are the highest of any region in the state (Reference 15).
Sleet or freezing r ain can occur during the colder months of the year where rai n fall s through a very shallow layer of cold
- 2. 3 -8 REVISION 9 - DECEMBER 2002
BYRON-UFSAR 33 . "Atmospheric Dispersion Code System for Evaluating Accidental Radioactivity Releases from Nuclear Power Stations; PAVAN, Version 2, Oak Ridge National Laboratory," U.S. Nuclear Regulatory Commission, December 1997.
- 34. Regulatory Guide 1.145, Atmospheric Dispersion Models for Potential Accident Consequence Assessments at Nuclear Power Plants," Revision 1, November 1982.
- 35. Regulatory Guide 1.23 (Safety Guide 23), "Onsite Meteorological Programs, February 1972.
- 36. NUREG/CR-6331, "Atmospheric Relative Concentrations in Building Wakes," Revision 1, May 1997 (Errata, July 1997).
- 37. Regulatory Guide 1.194, Atmospheric Relative Concentrations for Control Room Radiological Habitability Assessments at Nucl e ar Power Plants," June 2003.
- 38. NUREG/CR-2919, "XOQDOQ: A Computer Program for the Meteorologi c al Evaluation of Routine Releases at Nuclear Power Stations," Final Report, September 1982.
- 39. "Atmospheric Dispersion Estimates in the Vicinity of Buildings," J. V. Ramsdell and C. J. Fosmire, Pacific Northwest Laboratory, 1995.
- 40. Regulatory Guide 1.23, Revision 1, Meteorological Monitoring Programs for Nuclear Power Plants, Revision 1, Ma rch 2007.
41 . Desi a n Analysis ARA-0 0 2116, Tornado Missile Analysis of
{Byron Generating Station) BGS , Revision 3 2 . 3 - 52a REVISION 15 - DECEMBER 2014
BYRON-OfSAR 3.5.5.2.3 TORSCR TOP~3CR .i..s a FORTRAN c:omput.er c:cicle t.h~i t is used to post-process T1)!<.M1::; output filr::s. IU.i primary function is to compute Boole<-:tr1 corr~inations of target hit ancl damage probabilities over multiple t.arqets.
3.5 . 5.2.4 LS-DYNA LS-DYNA is a nonlinear explicit finit~ r::lement code for the dynamic analysis of structures. Since 1987, the LS-DYNA code has been extensively developed and supported by the Livermore Software Technology Corporation and is used for a wide variety of crash , blasl:. :;incl impact applications . LS-DYNA was lJSed to develop missile thresl1ol d damage velocities for selected targets which are used as an input in the TORMIS model .
3.5.5.J Analvsis The Byron TORMIS tornado missile risk analysis results show that the c=i.Li.thrnet.ic sum of damag;:, frequr::ncies for E1ll tart,Jet -:iroups affecting the individual units (i . e . , Unit 1 plus conunon components and Unit 2 plus *:*)Itunon components) are lower than the acceptable threshold frequency of l . OE-06 per year established in SRP Section 2.2.3 and Reference 21 .
The following limiting inputs and assumptions wE:re us*?.d in the analysis (refer to Reference 15 for e:-dditional assumptions and engine8ring judgments used in the analysis):
- a. A site specific tornado hazard curve and data set for Byron wa0 developed using statistical analysis of the t*10.r\..!\./Nationa1 Weather S*?.rvice Storm Predic:tion (>?.nter tornado data for the years 1950 thru 2013. The analvsis utilizes the Enhanced Fu"ita (EF) Scale winds eeds in the TOPMIS simu ations.
- b. The missile characteristics and locations are based on a plant walk down survey and plant drawings. Thr:: plant walk down survey was performed during a unit outage to capture both non-outage and outage conditions during the survey.
A stochastic (time-dependent) model of the missile population is implemented in T()FMIS . The stochastic approach to the missile population varies the missile populations in each of the TORMIS replications to account for predictable changes in plant conditions (i.e . ,
increased missiles during outages) and the randomness inher8nt in the total number of missiles present at the plant at any given time.
- c. Finite element cal~ulations were performed to provide the missile damage threshold velocity for each missile type to c.::iuse unacceptable crimping damaqe for the S:<C:T riser pipes, diesel driven auxiliary feedwater pump e~haust pipes and cover plates and the main steam power *=*pE:rated r0lief valve tailpipes.
3.5-2Gc
ATTACHMENT 1-2 Updated LAR Table 5 BYRON STATION UNITS 1AND2 Docket Nos. 50-454 and 50-455 Renewed Facility Operating License Nos. NPF-37 and NPF-66
~:!
9 2 ~ 2 11'1 2 2 0
9 2 ~ 2 ~ 0
~ l;I 2 ~ ~ !'!
0
!! I~ ~ i.:;
~ 8 :'? r: **,;, .VJ w Yi t~
<> P.: ~:
~
I~ ;'?
~
~
8
'(> on ~
9 Ill c.
~ .en~. ..en~ ...-.~
- I 0
CJ - - - - - - - - - - - - - ' - - ! - - -
....CIJ .'
- 1- -1...- - - - - - ~'--*---l-'-!l-'-'-+--L-- *- - - - -
t*
...coCl *~
..,..::: ~.
~
~ ~., ... ...
I-C: '
~
,_*"~ ,_*"~ ,_~ ,_~
0 ~
- ~
en co ~
c e
Ill Ill .2
-Cll-co...
.c co Cll I->
...Cll
- c.
u c:
CIJ
- I tT u.
Cll Cll Cl co E
l1' 0
c:
co Cll
- e