ML20199A017
| ML20199A017 | |
| Person / Time | |
|---|---|
| Site: | Millstone  |
| Issue date: | 12/16/1997 |
| From: | Harper G, Jeffrey Jacobson, Ohara T NORTHEAST NUCLEAR ENERGY CO. |
| To: | |
| Shared Package | |
| ML20199A002 | List: |
| References | |
| NUC-176, NUC-176-R01, NUC-176-R1, NUDOCS 9801270090 | |
| Download: ML20199A017 (30) | |
Text
. _ _ _ _ _ _ . _ . ._.
ORIGINAL PAGE 1 of 21, PAGES Rev. 1: PAGE 1 of PAGES Pev. 2: PAGE 1 of PAGES Rev. 3: PAGE 1 of PAGES QA RECORD 7 IMS NO. _NUO 01.16'.03 g_, YES RECORD TYPE 07.C16.004
,,,,, NO W.O./P.O. NO. 9398
- with 3 Attachments for a total of 32 pages YANKEE NUCLEAR SERVICES DIVISION CALCULATION / ANALYSIS FOR TITLE Diesel Oenerator Tornado Missile Evaluation PLANT M111 sten d CYCLE N/A CALCULATION NUMBER NUC-176 PREPARED BY REVIEWED BY APPROVED BY SUPERSEDES
/DATE /DATE M DATEh CALC./REV. NO.
obso !
REVISION 1 !
REVISION 2 REVISION 3 KEYWORDS TORNADO / MISSILES. RISK Attachment Pages 1 2 2 6 2 3 FORM WE-103-1 Revision 2 h
P
Page 2 of 21 Millstone Unit 3 NUC-176 Diesel Generator Tor.r 4 issile Evaluation 12/12/97 Prepared By: C#_ ud - .
Reviewed By: f TABLE OF CONTENTS Section Page 1.0 Calculation Objective and Plant Walkdown 3 1.1 Calculation Objective .3 1.2 Plant Walkdown 3 2.0 Method of Solution 5 2.1 Methodology 5 2.2 Precedence for Use of Methodology 6 3.0 Development of inputs and Assumptions 7 3.1 Target Descriptions & Drawings 7 3.2 Roof Exhaust Stacks (Targets 5 or 6) 7 3.3 EGE Wall Exhaust Piping (Targets 1 A or 2A) 9 3.4 Diesel Exhaust Stack Target Area 9 3.5 EGE Wall Exhaust Plenum Piping Area 10 3.6 Conditional Probability of Missile Penetration 11 3.7 Joint Probability of Missile Impact 11 4.0 Missile Impac./ Damage Calculations 13 4.1 Missile Impact / Damage Calculation (Targets 5 or 6) 13 4.2 Missile Impact / Damage Calculation (Targets 1 A or 2A) 13 4.3 Missile Impact / Damage Calculation (Diesel A or B) 13 4.4 Combined Missile Impact / Damage Calculations 14 4.5 Sunsnary of Results 14 5.0 NRC Tornado Missile Damage Acceptance Criteria 16 6.0 Conclusion 17 7.0 References 19 WE-103-2 20
-WE-103-3 21 Attachments Eages 1 2 2 6 3 3
Page 3 of 21 Millstone Unit 3 NUC-176 Diesel Generator Tornado Missile A. + tion 12/12/97 Prepared By: COW Reviewn - iy:._ SaA- .
LO CALCULATION OBJECTNE Ado PLANT WAIXDOWN 1.1 CALCULATION OBJECTIVE Attachment I contains a copy of a nquest for additional information (RAI) regarding the Millstone Unit 3 emergency diesel generator (EDG) exhaust stack.
As described in Question 1 of the Rit.1, a change was implemented in the last amendment (dated March 1996) to the FSAR to delete the licensipg commitment which required Northeast Nuclear Energy Company (NNECO) to open the EDG exhaust piping access hatches during a tornado alert. The purpose of the exhaust hatch is to provide an alternate exhaust path for the diesels in the event that the portion of the exhaust stacks downstream of the access hatch are significantly damaged by tornado missiles. Elimination of the commitment was based upon a site-specific probabilistic risk assessment described in Reference I which concluded that the probability of significant damage to an EDG exhaust piping from tornado generated missiles is less than 104 per year. The staff concluded (RA!) that the Reference 1 study only captures the risk of tornado generated missiles entering the exhaust plenums (roof and east wall openings) and does not address the probability of tornado generated missiles hitting EDG exhaust stacks on the roof of the EDG building. Therefore, the staff requested that a PRA be performed to demonstrate the risk of tornado generated missiles which could cause damage to any portion of the EDG exhau'st stacks. The PRA should consider all potential modes of tornado missile damage (i.e., impact of the exhaust stacks) not just missiles entering the
' plenum openings._ Therefore, the purpose of this calculation is to enhance the Reference 1 study by estimating the annual probability of tornado generated missiles impacting portions of the exhaust stacks which are vulnerable to tornado generated missiles.
1.2 PLANT WAIXDOWN On November.21,1997 a walkdown of the emergency generator enclosure (EGE) building was performed. The area around the EGE building was clean and free of sny construction material which could be a source of tornado missiles. Attachment 2 contains several pictures of the plant which were taken during the walkdown.
1
Page 4 of 21 Millstone Unit 3 NUC-176 l
-Diesel Generator Tornado Missile Evaluatio' 12/14/97 Prepared Byr7/$/ Reviewed By: cdL O L ,
Figure 1 is a view of the EGE building from Building 475 looking east. The diesel ;
exhaust stacks can seen on the east end of the EGE h'2ilding. Figure 2 is a close up view of a diesel exhaust stack. As can be seen, the exhaust stack is centered in the roof exhaust plenum and represents a relatively ameli target. The exhaust stack steel piping was measured to be % inch thick. Portions of the exhaust plenum around the exhaust stack are covered by bird screen and chalu. link fence. Figure 3 is a view .
- of the exhaust stack on the northeast side of the EOS building looking north. Figure -
4 is a view of the exhaust stack on the southeast side of the EGE building icoking to the east. Figure 5 is a view of the exhaust stack plenum on the east wall of the EGE building, These openings are covered with both bird screen and chain link fence. The exhaust gas piping for each diesel is located in the far left and far right sections of each EGE wall plenum opening shown in Figure 5. Figure 6 is a view looking out from the left side exhaust stack plenum shown in Figure 5. As can be seen, the exhaust stack piping is covered with rigid glass insulation (about 5" thick).
The exhaust stack piping is covered with this insulation up to the point where the exhaust stack exits the roof (Figure 2). It was noted during the walk down that missile entrance into the exhaust plenum on the east wall of the EGE building is a conservadve measure of potential tornado missile impact into the exhaust stack piping. -
The plant walkdown was performed by the preparer and reviewer of this calculation and by a NNECO engineer familiar with this issue.
4 e
i--- - , . . , - . - - - - ._,n,._,,.,,._n,,_, . , .- _ . , . , , , , , . . . , mm u --,,. - , . - - , , , , - , . - - , , - . , - - , . , . . - ., --
Page 5 of 21 Millstone Unit 3 NUC-176 ^
Diesel Generator Tornado Missile Evaluation 12/16/97 Prepared By: N Reviewed By: f d ,
2.0 MRTHOD OF SOLUTION 2.1 METHODOLOGY Tornado missile strike probabilities are typically estimated using Monte Carlo simulation techniones developed by EPRI (References 2 and 3). These analyses require detailed modeling of the number, type and location of missiles, tornado movement and windfield, missile injection and aerodynamic characteristics, and missile interaction with the plant. This methodology was used in the Millstone site-specific tornado missile evaluation (Reference 1). Use of the site-specific results to estimate the annual probability of tornado missile impact to a new target at the same site is documented in Addendum No. I to Reference 4, which was a detailed tornado -
missile analysis at the Seabrook nuclear power plant. Based on the Seabrook site-specific results, an evaluation was also performed to estimate the annual probability of a tornado missile impacting the exposed diesel exhaust stacks at the Seabrook site.
-Two modeled targets in close proximity to the exhaust stacks were used to estimate tornado missile impact density at the exhaust stack. Based on the evaluation, an SER (Reference 5) was written by the NRC which concluded that the expsed diesel generator exhaust stacks satisfy the Standard Review Plan Guidelines for tornado missile failure probability. The issues being addressed in this calculation parallel the concerns addressed in the Reference 5 SER.
Because the exhaust stack on the roof of the EGE is at essentially the same location as the two targets modeled in Reference 1, the tornado missile density can be estimated based on these site-specific results end used to estimate the annual probability of a tornado missile impacting the exhaust stack on the roof. Use of
- these site-specific results calculated in the 1985 Reference 1 study are conservativo for the following reasons:
1.- - The original computer simulations (Referece 1) reflected a mcxiest construction effort at Unit 3 resulting in a conservative missile population at the plant. Based on a November 21,1997 walkdown of the EGE building there are no construction lay down areas near the EGE building to provide a source for tornado generated missiles.
^ ~
p 6 of 21 Millstone Unit 3 'UC-176 Diesel Generator Tornado Missile Evaluation !.2/12/97 Prepared By:(dN Reviewed By: 94A - 1 r Therefore, it can be concluded that the number of construction l associated missiles used in simulations in the Refm:nce 1 study
- significantly exceed the number of construction associated missiles currently available at the site.
! 2. The machine shop and roof of the machine shop represented a significant source of tornado generated missiles in the original analysis.
The machine shop roof is directly east of the wall opening oa the east side of the EGE building. Based on the walkdown, construction l material is no longer on the roof of the machine shop and therefore the i
original analysis can be considered conservative, i .
2.2 PRECRDENCE FOR USE OF MRTHODOLOGY i
The NRC has issued a Safety Evaluation Report ((SER), Reference 6) on the EPRI L methodology (References 2 and 3). The SER concludes that the EPRI methodology
- j. can be utilized when assessing the need for positive tornado missile protection for j specific safety-related features in accordance with the criteria of SRP Section 3.5.1.4. Reference 4 describes a tornado missile evaluation of Seabrook Station l b'ased upon use of the EPRI methodology. In that evaluation the probability of missile impact that could lead to unacceptable damage (potential release of radioactivity) was estimated to be about 1.2 x 104 per year. The NRC issued an SER (Refemce 7) on the Seabrook submittal and concluded that the probability of
- unacceptht ilssile damage indicated in the applicant's analysis is conservative and l that the tre probability is likely to fall within the range of 104 to 104per year, thereby satisfying the SRP guidelines for tornado missile failure probability. Use of the results for the 30 targets analyred in the Reference 4 study was applied to -
estimate missile strike probabilities for other targets at Seabrook. The NRC'also issued an SER (Reference 5) on a tornado missile evaluation of the exposed diesel
- generator exhaust system at Seabrook Station. The conclusion of the SER was that the mean probability of a tornado missile impacting the diesel generator exhaust was 104 per year which was acceptably low and that hardened tornado missile protection was not required.
--,,.3,- ,---.my - -
.,~,<%-.,. -,,e 7- w-,-4-, ..~,, _ . 4., , ,c ,_,y ,,,-,,,----..,.-.-,.,-,,,,y 3 , , , . ,,m.v. -,,-r--
Page 7 0121 .
Millstone Unit 3 NUC-176 Diesel Generator Tornado Missile Evaluation 12/12/97 Prepared By:6 CW Reviewed By:_V d .
3.0 DEVE! OPMRNT OF INPUTS AND ARRUMPTIONS 3.1 TARGET DFRCRIPTIONS & DRAWINGS The staff requested that a PRA be performed to demonstrate the risk of tornado generated missiles which could cause damage to any portion of the EDG exhaust stacks. Reference 1 documents a site-specific probabilistic tornado missile impact evaluation at Millstone Unit 3 for openings in the EGE building. Four targets were modeled in this analysis. Targets 1 and 2 are the 48-in x 168-in plenum openings on the east wall of the EGE building at elevation 51' (S&W. DWG. No.12179-EB-7B-12). The diesel exhaust piping is inside these openings. Referenced drawings can be found in Attachment 3. Targets 3 and 4 are the 66-in. x 100-in, exhaust plenum openings on the roof of the EGE building (S&W. DWG. No.12179-EB-7B-12).
The diesel exhaust stacks (targets 5 and 6), which were not modeled in the Reference 1 study, are centered in the target 3 and 4 openings and extend up 3' above the exhaust plenum opening (S&W. DWG. No.12179-EB-7B-12). Exhaust piping thickness (0.5 inches) is shown on Vendor Drawing 2280.000-450. For one '
diesel exhaust system, tazgets 1,3, and 5 would represent the opening or actual targets vulnerable to tornado missiles, and the other would be represented by targets 2, 4, and 6.
3.2 ROOF' EXHAUST STACKS (TARGETS 5 OR 6)
To estimate the probability of a tornado missile impact on the exposed diesel exhaust stack (either target 5 or 6) th tornado missile impact probabilities presented in Reference 1 (Table V-1) for targets 1 and 3 will be adjusted by target area ratios.
The terget area ratio is the ratio of the exposed diesel generator exhaust stack area to that of the targe: 1 and 3 modeled target areas. The adjustment of the tornado missile impact probabilities by a ratio of the areas is consistent with the Reference 1 analysis. In the Reference 1 analysis each of the target areas were increased by a factor to increase the likelihood of a tornado missile strike to provide better statistical estimates of missile strike probabilities. The missile strike probability is subsequently reduced by this factor. haplicit in this approach is the condition that tornado missile impact probabilities are approximately proportional to the area of k
I Page 8 of 21 Millstone Unit 3 NUC-176 Diesel Generator Tornado Missile Evaluation 12/16/97 f
Prepared By:,A MM Reviewed By: hA- .
I the target at given locations relative to sources of missiles, shielding, target orientation and other factors that effect tornado missile impact probabilities.
From Table V-1 in Reference 1 the probability of a tornado missile impacting the modded EGE wall opening is 9.0 x 104 per year. The modeled EGE wall opening as described in Table IV-10 is 10' x 21.5' (215 ft2 ). The actual opening as 2
described in Table IV-10 is 4' x 14' (56 ft ). As described in Reference 1, the target planar areas of the openings are increased so that the modeled target openings in TORMIS are larger than those actually at the plant. The larger size will collect more hits for the same number of missile simulations and therefore provide an-estimate of the missile flux in the immediate vicinity of the opening. An adjustment '
factor K, is used to adjust the impact probabilities for the actual target size.
K, = T/T, (1) where T = actual opening target area and T, is the modeled target opening area.
Therefore, the adjusted multiple missile impact probability can be defined by P" = P" (A) K, (2) where P,"is the adjusted multiple missile impact probability of hitting the actual target, P" (A) is the multiple mistile impact probability of hitting the modeled target, and K, is the adjustment L ctor.
As can be seen from the above, the factor K, is simply a ratio of the actual target area to the modeled target area. Therefore, given the rodeled multiple missile impact probabilities from Table V-1 for both targets 1 and 3, estimates of the multiple missile impact probability for the exhaust stack can be calculated using the following equation P" = P" (A) - (exhaust stack area /modeled target area) (3) where P, " is the multiple missile impact probability for the exhaust stack.
Page 9 of 21 Millstone Unit 3- NUC-176 Diesel Generator Tornado Missile Evaluation 12/14/97 Prepared By: TdM Reviewed By: V A 3 3.3 EGE WAT T. EXHAUST PIPING (TARGETS 1 A OR 2A1 l
The targets associated with the exhaust piping inside the EGE wall exhaust plenums will be designated as targets 1A and 2A to avoid confusion with the results from Reference 1.
Table V-1 in Reference 1 describes the adjusted multiple missile entrance probability for the ECE wall openings. The entrance probability is adjusted for actual target area and missile size. This is described in Equation 6 from Reference 1, P," r= P" (A) K; K,,, ,
(4) where P,"is the adjusted multiple entrance probability for the actual target area and K,iis an adjustment for missile size in the prediction of first opening entrance probability.' P," describes the probability of multiple missiles entering the wall opening. Impact of the exhaust piping inside the exhaust plenums on the east wall of the EGE building is bounded by the multiple missile entrance probability of 6.6 x 104 described in Table V-1 of Reference 1. An estimate of the multiple missile impact probability on the exhaust stack piping inside the EGE wall exhaust plenum is provided again by Eqn. 3 with the exception thatN P (A) is replaced by P " and s the ratio is the approximate frontal exhaust pipe area to the actual wall opening area.
P" =
P," - (exhaust pipe area / actual wall target area) (5) 3.4 DIFREL EXHAUST STACK TARGET AREA Drawing ( S&W. DWG No.12179-EB-7B-12) shows the dimensions of the exposed diesel generator exhaust stack on the roof of the EGE building. The exhaust stack extends 3' above the protected opening and the top is miteced at a 30 degree angle.
Each exhaust stack is 40" (3.333') in diameter (S&W. DWG No.12179-EM-13B-10).
l The target area of the exhaust stack is described by the surface area of the 1
Page 10 of 21 Millstone Unit 3 NUC-176 Diesel Generator Tornado Missile Evaluation 12/14/97 Prepared By: I8/M Reviewed By: I A
- exhaust stack. The surface area is conservatively modeled as a 3 foot high cylinder with a 40 inch (3.33') diameter.
Target area = nDH where D = diameter = 3.333' and H = height = 3'. '
The target area = (n) (3') (3.333')
1
= 31.42 ft2 3.5 EGE WALL EXHAUST PLENUM PIPING AREA Missiles that enter the EGE wall exhaust plenum opening will not always directly impact the exhaust piping. As shown in View 2-2 of S&W DWG. No.12179-EM-13B-10, the exhaust piping in the exhaust plenum occupies less than 1/6 th of the EGE wall opening. As a first approximation, use of an exhaust gas piping target area which is 1/6 th of the EGE wall exhaust plenum area (56 ft2) is reasonable and consistent with area adjustment methods. Hov/ever, as a conservatism, the exhaust gas piping target area vulnerable to direct tornado missiles hits is modeled to be 1/3 of the EGE wall exhaust plenum area, or 18.7 ft2, 3.6 CONDITIONAL PROBABILITY OF MISSILE PENETRATION Table V-1 in Reference 1 contains perforation damage probability estimates. This is described as the probability per year of tornado missile perforation of a 0.2 inch steel plate located at the opening. As described in Reference 1, "The perforation damage criterion provides a conservative measure of the frequency of missiles with any chance of damaging the diesel generator exhaust piping. Since the exhaust piping is 0.5 in thick, the use of a 0.2 in thick flat plate provides a very conservative damage criterion." Based on the Table V-1 results, an estimate of the conditional probability of missite perforation given missile impact on a target can be
Page 11 of 21 Millstone Unit 3 NUC-176 Diesel Generator Tornado Missile Evaluation 12/12/97 Prepared By:IOW Reviewed By: GL- __ .
I determined.
For target 1 the impact probability is 9.0 x 104. The perforation probability is 6.2 x 10. Therefore, the conditional probability of perforation to the exhaust piping given missile impact is estimated to be (6.2 x 10~7/ 9.0 x 104) = 0.07. As described above, use of this factor is reasonable given that the exhaust piping is 0.5 inches thick.
3.7 JOINT PROBABILITY OF MISSILE IMPACT To combine the probabilities of various results the following axiom from probability will be used:
P(AuB) =
P(A) + P(B) - P(AnB) (6) where P(AuB) is the probability of event A or B or both, P(A) is the probability of event A, P(B) is the probability of event B, and P(AnB) is the probability of event A and B.
- The event of interest is when the vulnerable exhaust piping for both diesels experience tornado missile impacts given a tornado. This event is described as the joint probability of the two events, or P(AnB). In order to assess the reduced probability associated with the intersection of these two events the results from Table V-1 of Reference 1 are used. The impact to either Exhaust A or to Exhaust B (P(AuB)) is 1.2 x 104. Impact to Exhaust A and Exhaust B (P(AnB)) is 6.8 x 104. Based upon these results, the factor to adjust the union event probability to an intersection event probability is (6.8 x 10-8/1.2 x 104) = 0.057.
Page 12 of 21
+Itil ae Unit 3 NUC-176
'iec Generator Tornado Missile Evaluation 12/12/97
. Prep M By:7dM' Reviewed By: e=A - 3 0
4.0 MISSTI R IMPACT' DAMAGE PROBABILITY CALCULATIONS 4.1 MISSILE IMPACT / DAMAGE CALCULATION (TARGETS 5 OR 61 The annual probability of tornado missile impact to either the target 5 or 6 diesel exhaust stack on the roof the EGE building is calculated based on use of Eqn. 3 (Section 3.2), the exhaust stack area (Section 3.4), and the unadjusted impact probabilities from Table V-1 in Reference 1.
P." = P" (A) - (exhaust stack area /modeled target area) where the exhaust stack area = 31.42 ft2 , the modeled target area = 215 ft 2, and P"(A) associated with the EGE wall opening = 9.0 x 104 .
P" =
9.0 x 104 (31.42/215)
P" -
1.3 x 104/ yr.
Use of P" (A) associated with the roof opening = 7.1 x 10* results in an impact -
probability of about 1.0 x 104 Use of the roof opening impact probability is unconservative'in that it primarily accounts for vertically traveling missiles.
The annual probcbility of missile perforation to either targets 5 or 6 given tornado missile impact, based on the adjustment factor developed in Section 3.6, is (1.3 x 104) - (0.07) = 9.1 x 104 .
Targets 5 and 6 are centered in the EGE roof exhaust plenums which are modeled as targets 3 and 4 in the Reference 1 study. Vertically falling missiles that could impact or plug the exhaust stack are accounted for by the target 3 and 4 exhaust plenum opening results. The impact probability result for either target 3 or 4 from Table V-1 in Reference 1 is 1.5 x 10*.
c _ _ _ _ _ - _ _ _ - _ _ _ _ _ _
Page 13 of 21 Millstone Unit 3 NUC-176 Diesel Generator Tornado Missile Evaluation 12/12/97 Prepared By:. [ 88 Reviewed By: 7aA - .
4.2 MISSII R IMPACT / DAMAGE CALCULATION (TARGETS 1 A OR 2A)
The annual probability of tornado missile impact of the target 1A or 2A exhaust piping in the exhaust plenum on the east wall of the EGE building is calculated based on use of Eqn. 5 (Section 3.3) and the exhaust stack area determined in -
' Section 3.6.
P ," = P," - (exhaust pipe area / actual wall target area) where the exhaust piping area = 18.7 ft 2and the actual wall target area =
56ft.2 P ,N -=
6.6 x 10-7 (18.7/56)
P,," = 2.2 x 10-7 / yr.
Consistent with the NRC request, this probability reflects the probability of tornado missile impact to portions of the exhaust stack piping vulnerable to tornado missiles.
The annual probability of missile damage to targets 1A or 2A, based on factors developed in Section ?.6, is (2.2 x 10-7) - (0.07) = 1.5 x 104 4_3 MISSIT E IMPACT / DAMAGE CALCULATION (DTERFI A O1L]D For a given diesel, tornado missile impact can occur to the exposed exhaust stack on the roof, through the EGE wall opening, or through the roof exhaust plenum opening. Impact to the exhaust stack on the roof is estimated in Section 4.1 to be 1.3 x 104/ yr. Impact to the exhaust stack piping through the EGE wall is estimated in Section 4.2 to be 2.2 x 10-'/ yr. Impact through the roof exhaust plenum from Table V-1 in Reference 1 is 1.5 x 104. The total tornado impact probability per year for a given diesel can be estimated by summing these individual impact probabilities. Summation of these impact probabilities to obtain the total tornado impact probability per year is conservative because it does not account for the
Page 14 of 21 Millstone Unit 3 i TUC-176 Diesel Generator Torr.do dissile Evaluation 12/14/97 Prepared By:7 M __ Reviewed By: 4ak ,
/
combinations ofjoint probabilities associated with the three impact probabilities.
Therefore, the total torna.'. impact probability p:r year for a given diesel is 4
(1.3 x 104 + 2.2 x 10 + 1.5 x 104) .= 1.5 x 104 .
The annual probability of missile damage to a given diesel is based on factors developed in Section 3.6. Therefore, the total tcInado damage probability per year for a given diesel is estimated to be (1.5 x 104 ) - (0.07)= 1.1 x 104 .
4.4 COMBINED MISSILE IMPACT / DAMAGE CALCULATIONS Summatien of the impact probabilities to obtain the total tornado impact probability per year to either diesel A or B is conservative. Therefore, the total tornado impact probability per year L either diesel A or B is estimated to be 3 x 104 This .
probability represents tne annual probability of either the diesel generator A or B portions of exposed exhaust piping being impacted by a tornado generated missile.
The joint probability of both the diesel generator A and B portions of exposed exhaust piping being impacted by a tornado generated missile given a tornado is estimated based upon the adjustment factor developed in Section 3.7. Therefore, the mean annual probability of tornado missile impact to both the diesel generator A and B portions of exposed exhaust piping is (3 x 104) - (0.057) = 1.7 x 104 .
The annual probability of missile damage to both diesels is based on factors developed in Section 3.6. Therefore, the tornado damage probability per year for 4
both diesels is estimated to be (1.7 x 10 ) - (0.07)= 1.2 x 104 .
4.5
SUMMARY
OF .RESULTS The following table summarizes the tornado impact probabilities for each of the targets described above. Targets 1A and 2A are the diesel exhaust piping inside the EGE east wall exhaust plenum. The target 1A and 2A probabilities have been adjusted from the Reference I study in that impact and perforation probabilities are associated with the diesel exhaust piping and not simply the wall opening. Targets 3 and 4 are the roof opening exhaust plenums. The probabilities associated with 1
s Page 15 of 21 Millstone Unit 3 NUC-176 Diesel Generator Tornado Missile Eve.' .ation 12/12/97 Prepared By:fe# d Reviewe/ By: # 3 these targets come from the Reference 1 study. Targets 5 and 6 represent the- _
exposed' diesel exhaust stacks os te roof of the EGE building. The exposed exhaust piping for diesel generator A-(DG-A) is described by targets 1 A, 3, and 5.
The exposed exhaust piping for diesel generator B (DG-B) is described by targets 2A,4, and 6. -
TABLE 1
SUMMARY
OF RESULTS Target Number Target Group Impact Perforation ProbabilityNr. ProbabilityNr. -
1A,2A Exhaust Pipes 2.2 x 10-7 1.5 x 104 Inside East EGE Wall Openings 3,4 Roof Plenum 1.5 x 10* -
E*haust Openings 5,6 Exposed Roof 1.3 x 104 9.1 x 104 Exhaust Stacks 1Av3u5 DG-A 1.5 x 104 1.1 x 10-7 2Au4u6 DG-B 1.5 x 104 1.1 x 10-7 Either DG-A or 3.0 x 104 2.2 x 10-7 DG-B Both DG-A and 1.7 x 10 1.2 x 104 DG-B
4 Page 16 of 21 Millstone Unit 3 NUC-176 Diesel Generator Tornado Missile Evaluation 12/12/97
- Prepared ByCO'# Reviewed sy: saA -
p ,
5.0 NRC TORNADO MISSIT R DAMAGE ACCElrrANCE CRITERIA -
NRC licensing criteria concerning tornado missile protection are provided in SRP Sections 3.5.1.4 and 3.5.2. These criteria specify the tornado missile protection requirements for safety-related systems. SRP Section 3.5.1.4 also provides guidance on the use of probabilistic risk assessment (PRA) methodology in lieu of a deterministic approach for assessing tornado missile protection. The probabilistic acceptance criteria for tornado missiles is the same as that identified in SRP Section 2.2.3.
The probabilistic acceptance criteria from SRP 2.2.3 is as follows: "...the expected rate of occurrence of potential exposures in excess of the 10 CFR Part 100 guidelines of approximately 104 per year is acceptable if, when combined with reasonable qualitative arguments, the realistic probability can be shown to be lower."
References 5 and 7 describe the use of PRA methodology in the assessment of tornado missile damage to systems that could cause a release of radioactivity in excess of the 10 CFR Part 100 limits. These references also document the acceptability of the approc' with the NRC. Due to-the acceptably low risk, hardened tornada missile protection was not required at Seabrook Station for some safety-related items. Tornado missile impact to a given diesel generator exhaust 4
stack at Seabrook was estimated to be about 10 / year. Given this result, an SER was issued by the NRC (Reference 5), and hardened tornado missile protection was not required. Similar conclusions should be appropriate for application at Millstone for the existing diesel generator system.
l Page 17 of 21 Millstone Unit 3 NUC-176 Diesel Generator Tornado Missile Evaluation 12/16/97 Prepared By:'
Reviewed By: ( ,
6.0 CONCLUSION
An analysis was performed to estimate the annual probability of tornado missile impact on the diesel exhaust piping which is potentially vulnerable to tornado generated missiles. For a given diesel, the exposed roof exhaust stack, roof exhaust plenum opening, and exhaust _ pipes within the EGE east wall opening represent portions of the diesel exhaust piping that are vulnerable to tornado missiles. Based on this analysis the annual probability of tornado missile impact on portions of a-given diesel exhaust system which are vulnerable to tornado missiles is estimated to be about 104/ year. Inherent in this estimate are the following-conservative-assumptions:
- 1. The number of potential missiles present during construction and used in the Reference 1 analysis are no longer present,
-2. The exposed roof exhaust stack was conservatively modeled,
- 3. Engineered structures in the Reference 1 analysis were conservatively-modeled as failing like non-engineered structures,
- 4. - The tornado missile impact probability for a given diesel is about 104/ year. Using a conservative representation of damage (0.2 inch thick steel to represent 0.5 inch thick steel) the damage probability for a given diesel is estimated to be about 10-7/ year.
- 5. Damage to both diesel exhaust systems given a tornado is estimated to be about 104/ year.
These results are consistett with the results described in the Reference 5 SER which did not require hardened tornado missile protection for the exposed diesel exhaust stack at Seabrook.
Based on the above and acceptance criteria described in Section 5.0, it is concluded that the need to provide tornado missile protection for the limited portions of each
hp 18 of 21 Millstone Unit 3 NUC-176 Diesel Generator Tornado Missile Evaluation 12/12/97 Prepared By:7dB Reviewed By: v 9a b ,
?
diesel exhaust system that are vulnerable to tornado missiles is not required. In particular, it has been shown that the probability of the initiating event that could lead to potential exposures in excess of the 10 CFR Part 100 guidelines is on the order of 104, and when combined with the conservatisms described above, the realistic probability can be shown to be Icwer.
In summary, the probabilistic acceptance criteria for a relcase in excess of the 10 CFR Part 100 guidelines is canservatively met and hardened tornado missile protection for the limited portions of the diesel 2xhaust system vulnerable to tornado missiles is not reouired.
The objectives of this calculation have been met.
6
Page 19 of 21
illstone Unit 3 NUC-176 Diesel Generator Tornado Missile Evaluation 12/16/97 Prepared By:hO d Reviewed By: f ,
7.0 MEFERENCFR '
- 1. Tornado Missile Risk Analysis of Millstone Unit 3 Emergency Generator Enclosure Opening, Final Report C5945.3, Applied Research Associates, Inc., April 1985.
- 2. Tornado Missile Risk Analysis, Electric Power Research Institute, EPRI NP-768, May 1978.
- 3. Tornado Missile Simulation and Design Methodology, Volume 1: Simulation Methodology, Design Applications, and TORMIS Computer Code, Electric -
Power Research Institute, EPRI NP-2005. August 1981.
- 4. Seabrook Nuclear Power Plant Tornado Missle Analysis, Final Report C569, Revision 1, Applied Research Associates, Inc., March 1984'with Addendur.
No. I dated December 1984 and Addendum No. 2 dated December 1984.
- 5. NUREG-0896, Supplement 5, Safety Evaluation Report Related to the Operation of Seabrook Station, July 1986.
- 6. NRC Memorandum from L. Rubenstrin to F. Miraglia,
Subject:
" Safety Evaluation Report regarding EPRI topical reports, ' Tornado Missile Risk
. Analy*is' (EPRI NP-768 and 769) and ' Tornado Missile Risk Evaluation Methodology' (EPRI NP-2005), October 26,1983.
- 7. NUREG-0896, Supplement 3, Safety Evaluation Report Related to the Operation of Seabrook Station, July 1985.
EVALUATION OF COMPUTER CODE USE CALCULATION NO. .NUC-176 REVISION NO. _ Page 20 of 21 NUC-176 12/16/97 List the computer codes used, and complete the following: '
Approved Appropri-per ateness Outstanding WE-108' Verified 2 SPRs' Code Name/ Version Yes No Yes No Yes No N/A = no crxles used Refer to Section 4.1.4.4, Bullet 3, of this procedure.
3 Refer to section 4.1.4.4, Bullet 2, of this procedure.
8 Refer to WE-108, ssetion 4.4.
If a computer code was not verified per WE-108, or if there are outstanding SPRs, state below why it is appropriate.
Code Name Appropriateness N/A FORM WE-103-2 Revision 2
CALCULATION / ANALYSIS REVIEW FORM Page 21 of 21 NUC-176 12/16/97 CALCULATJON NO. NUC-176 REVISION NO. _
r COMMENTS RESOLUTION 0:4M lA fit f es t hre e f W ex .' & ( >> E - t C) % '
0+ cs W .'noe C OPM b e t ,x w h t vo c4 M M M b bt-eo<,t_c+ , J r-eso L.,4 '
\d o w i. - <. Lk V40.' m3 D r Ot P S.
(
Identify method (s) of review:
Calculatic.n/ analysis review O Alternative. calculational methoo O oualification testing Resolution By: ## %
Preparer /Date A2//(gy7 Comments Continued on Page Concurrence with Resolution s -
le a_ -
s.vi .ri..ag,qq FORM WE-103-3 Revision 4
Jt%4 ! ,
Mr. Neil S. Carns Senior Vice President [#9 //.2.
and Chief Nuclear Officer /Vue. - / 7 6 Northeast Nuclear Energy Company gg c/o Ms. Patricia A. Loftus Director - Regulatory Affairs P.O. Box 128 titerford. CT 06385 ,
SLR)ECT: RE0 VEST FOR ADDITIONAL INFORMATION RE(MRDING THE MILLSTONE UNIT 3 EMERGENCY DIESEL
. GENERATOR EXHAUST STACK (TAC NO. M97508)
Dear Mr. Carns:
By letter dated July 24.1997. the staff provided you an update of the NRC's Millstone Restart Assessment Plan (Plan). Enclosufe 3 to the Plan contained the Significant items List (SIL) for Millstone Unit 3. Sil #75 for Millstone Unit 3 addressed two inspector follow up items; one concerning the emergency diesel generator exhaust stack. The NRC staff has reviewed the closure package for this item and in order to support further NRC evaluation, you are requested to provide the NRC with answers to the enclosed que;tions. Please respond to this request for additional infonnation expeditiously in order for the staff to complete 1ts review in a timely manner.
Sincerely.
James W. Andersen. Project Manager Special Projects Office - Licensing Office of Nuclear Reactor Regulation -
.bcket No. 50 423 inclosure: As stated
- c w/ encl: See next page 0
_. . w. . ;------ : ::- . ---- . ; ,
p;. g RE0llFST FOR ADDITIONAL INFORMATION
/WA-[m 4 I /
CONCFRNJNG FMERGFNCY DIESE1 GENFRATOR EXHA!M SVACKS
/Vvt.~ /?6 MILLSTONE NilCLEAR POWR STATION llNIT 3 7g TAC NO. M97508 1.
Based on a probabilistic risk assessment (PRA) described in the final sumary report, fTornado Missile Risk Analysis of Hillstone Unit 3 Emergency Ccnerator Enclosure Openings.* dated March 1985, Northeast huclear Energy Company (NNECO) concluded that the probability of significant damage to an emergency diesel generator (EDG) exhaust piping from tornado generated missiles is less than 1 x 10 6 per year. Accordingly, a change was implemented in the last amendment (dated March 1996) to the Final Safety Analysis Report (FSAR) to delete the licensing comitment, which required NNECO to open the EDG exhaust piping access hatches during a tornado alert. The staff finds that the above cited PRA was only perfonned to demonstr6te the risk of tornado generated mistnes, which could enter any of the openings in the exhaust plenums to cause damage to the EDG exhaust system. It does not address the probability of the EDG exhaust stacks 1.ocated outside the building being hit by tornado generated missiles. Therefore, provide a PRA to demonstrate the risk of tornado generated missiles, which could cause damage to any portion of the EDG exhaust stacks. This PRA should consider all potential modes of tornado missile damage to the exhaust stacks, not just the probability that missiles will enter one of the openings of the exhaust plenum, '
!. Your risk assessment for protection against tornado generated missiles only addresses the acceptance criteria for a single vulnerabil_ity, i.e., the EDG exhaust system. In order to rely on probabilistic arguments that tornado missile protection is not reautred, the risk assessment (by must not EFoviding also showlthat protection the system) for the EDG exhaust overall beyondrisk from criteria the acceptance the guidance of Standard Review Plan (SRP) Section 2.2.3. which aeais with identiftcation of~
trsrnado discusse aenerated_
design basis events using probabilistic methods. Therefore, the analysis should also consider all other structures, systems, or components (SSCs) not protected from tornado generated missiles that are required to prevent a release of radioactivity in excess of 10 CFR Part 100 that are not protected from tornado generated missiles. This guidance states that an expected rate of occurrence of potential exposures in excess of the 10 CFR Part 100 guidelines of approximately 10-5 per year is acceptable if, when combined with reasonable qualitative arguments, the risk can be expected to be lower. This request for additional infonnation need not be addressed if all other safety-related SSCs (other than the EDG exhaust system) are adequately protected (by existing barriers / structures) from the effects of tornado generated missiles.
, , y T
gh . .
.y . , . . . .. .
- ' ~
J %
d $7
, 6 s
0 g* ..
.6 -
- s. e 4
~
!,y- 0
~
y ,..,
f'."
9 . . :, ,
, 3
.- ; w,.
. . r ,.
[ ,, i '. . " . . .
,.(, . ,.
~
, , t.; -
94 4, - . , .
I e
. i
~ _kT -
" I
[ gG $44 , * '} .
7
.s N-m .. ;-
is . < ,.
,. '. l.. ..-. .+ - 't'a .Q u cy . . . it .
- ,3 ,
h I
)
W 57'lt . ?'
JUL J ..- jM9 $I."... '
yn x.mL r -
uw., ~
%f '
? _
, :j ( ,
.L _4 y- -
(343. 3nhk %y ,d i g g M q. ,r. 7 :b1 . .
cia -
4,' e t;. . -
.g m 3;.g gs 23
- , > - -- . s y i _p -
GR. . a*J.,! .,t y.
s a ' ,, .Mh *
- i. +
. ,- .~ .
he h
. I ..!.'. '
,[ I kir[.hF , i el l- a a .
gL (@, ;I{, W ^ @$iO*,tf. _p1 , 4
. . , , n n- wa '
~;
fff?%.;
c 34 k w Yf Q s
' $w) ._b a t,
& i .+ , }W g";,![ _
jsYiW '
.a ,ip ait- $ ? k Y * ..
- <mQ
- a 'k ..% : 1 . ,;;al4c . .;.
r 4 t$r : .
3 } &., .. ' s
-'- -c%wQ %[( y, -
.c
?..: -
n bh' ,y y
- 3 "3(A 'b 1
,.;...;.['.%-
an w,, v.. .s+ .Y ) l ,,M .:-,%[. ,.
= - ,-
up c. , q%s.
- u it 30 -
, f* f ., :=.. '
'.w W ;
s
. ~
- ~
Yv 3 % , ,3 '
0
, g;4 ; - *
- If' g .= t ,
,d['
g .s w .
(.Ii .
d '*
n III q i ,
' ._. g - -
i- ,
.. j1 + ,
?. ftA<. . .c .
r- ,
S.l' . :.
acy., + -
l s l~ .g , ,y. .
. s O.
'7 ? _"
yka dg;, J t - --- .
' .e y; - -
gr c -
e
'i [T2 / d[
t u
'l :
. - 1
.n. /-
Mj ,q .n%j! /'/ om [ '
, .4 :,r- -y %m d
_m '4 w
-u$- ' Wu A \ 4 :N lY s- . ..
//K( , %3 '
7
. -a ;-
l.-
+>
s
)
p
\\
, Ik u p <r , 2 PJ i.
\{
pr, - '
1 ~ ~ ~ ~ . . . - - - -
/psg1;mq e c e f, ; ,;;
{-
,2 # 4 ",' jj
_ , .J
])
fx
- ;j
- i n
,1 d . 5, ' -
=; c i
rc%i ,
%;La 4 I gt r
{
. -I
. I,..'
i ~
kl -
) m c
x '
IE T_ <
7; _t j-
- f so k 5.,i t7tm i:. :{
y .
';gi
[
it ?
pl,
. *n C
f
~
y .
o,i i o 1 n*m e
g' m % 1i+t ' ;f y# .ll ( s* .m t,
' ~;
+
~ '!
igq)jji1-C- +
Y wj r 1
, , n ,
4 E [ww,$Wy=,$ kgh wgn
. wm -
t ;. ; '* b Y' c .,ws=
- p. s ..
jj # + y- 3 w:a g
2- 1 : 2 l4
- 'M r g;c[ 4 dtii y
X & - @- y;o y#u MI
,[b;g, .a v'll* p
.f* I' l ll- /4 Qd 'W i '"
.g ?
,) , ';q^-+ -
c!
'.. ~n s. " - '[.
W(gs.4j$dy J ,
- N.
Y, 's l1f,,,'
Njhj\ :, '
hklh$hll
~
j
- _e 8
z, e y . .;
L ._ _ - .__ A '*- > l(.4 g m' hA Lf -m fd n + . ~-8--- -
. l 4 .g f fII 4 g
}h .
-f{
k ("
. ).
4.
x l . .
.z
- t A .! _ y
- 1. c. . ' ; '
.s..
- q J af l
}l s
wlx[ :as N 2, 0 ft 9 'I d N' <
y Me -
}_ _ _ _ _ . - - -
$?-
- p* 1i // ,<'
. s =;. y 3:. %,
-O _
yN ,
~
N:_3 ppp '
.i Y_y?
e
.q pyg#
p* ( w r .
x%yc .
+
a
~- ,
! '^
.7.
e r, ((l* 3 '
ks 4 p^ 4j[
t q
g
' t- ,
[%f;zW ,%-\y Di , *
$.5e fy p 4 y,e d ?* r p
..m.
- k N. .'
O kj g:%~
R y
- f. '
,t.. .
.z ;,
$1*
- f. %
e
{. : ,' ,4 ~4- , y
{.
+. .3 e
4 4 .%i g -
' %~ ~
t e .f 4 3 --
jq :
sp L. < 1, -g* ..,g=* _
5 m
1 e (g
' .(.
%l ,
i
+, x, ; -
A n ,, .
}t .
y '
% )> ; W; } '
w 1
, E 4
. u. vn8 s
'g' s) s '
+$'--
I'
,> f. - ' .. t
'; ' ,k t. h ; . Q * \ . &
.- ..*sp1 3 .3 .g t.
.\,
- S
-}i .) -. ) . j
- W t-
? >
gasm,, Nt:# ;.
w].g.
' 4 0 t- y9 +
t g; '
(Ihf b ;t . q 1- -
s h ; ? , % . . (1
. .; dj0 k N I!g +
1 eAe 2q y
" r j l)3.y F m- ]FR WhW2_$k A ptt 2- i p
b
( &s [
~
M v;4, ,
4ys T SAI(Qi '
-y. -
>[
'A.'
', [ q
'N . *# I &
o . .
'5I
33hwg k p
.y,.-
k=
- a
- g. '
rX,+ :;\ .
La y T. ,
q.h
, w" *
} '
,, h
- f' ,
. . . y .. ~
((
b,. / * .
T .)*e u , b. -i,.. " - -
g C=
+g
~
.i.
r ~. -
- ~ ~ - -
f+;.
- %. . NT . 4 '
,m-g!
. j k
i i
.(
s um # .,
f
. . ' - gey ~ T*
- f 4$. ', .
T qa w _. >.
(+ e .
m ,~
- y . . , .. v
~g %,,1; ,.
,Q* %
... ~ , .- 'fp- ;. , L -
- ym .m:, u N 4 4; m- ' .-
ypp: - - 3* * #-
}'yyr*4; . .
(fM ? , , y E '
+
c._
~3 m J. iL q- y. _,. . ;
k:'Y hlhh+l '
$snm ja .
'.% !gg i*.)
- . l 'y-(;;9 ;. -t [?,
'W,
^13id j
- . ,1 y.
1
- 4. g, c ; .
['g ~
W--
. - $-_A 7 '.?o,- Y#'
^
c._' ; ' f .
U.S. Nuclur Regulatory Commission B16767\ Attachment 2\Page 1 Attachment 2 Millstone Nuclear Power Station, Unit No. 3 NRC Request for Additional Informatlon Regarding the Millstone Unit 3 Emergency Diesel Generator Exhaust Stack (TAC No. M97508)
Commitments January 1998
U.S. Nucber Regul: tory Commission B16767\ Attachment 2\Page 2 List of Regulatory Commitments The following tabic identifies those actions committed to by NNECO in this document. Please notify the Manager - Regulatory Compliance at the Millstone Nuclear Power Station, Unit No. 3 of any questions regarding this document or any associated regulatory commitments.
Commitnient Committed Date Submit response to Question No. 2 identified in NRC letter dated February 15,1998 September 15.1997 2
% g w w
OVERSIZE DOCUMENT PAGE(S) PULLED SEE APERTURE CARD FILES APERTURE CARD / PAPER COPY AVAILABLE THROUGH NRC FILE CENTER NUMBER OF OVERSIZE PAGES FILMED ON APERTURE CARD (S) e3 ACCESSION NUMBERS OF OVERSIZE PAGES:
9702020215 9 7 0 2 0 2 0 2 to
% o 9nCD'to-0)
DAP 08 U