ML20030B491
| ML20030B491 | |
| Person / Time | |
|---|---|
| Site: | Sequoyah  |
| Issue date: | 08/10/1981 |
| From: | Mills L TENNESSEE VALLEY AUTHORITY |
| To: | Adensam E Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 8108180193 | |
| Download: ML20030B491 (6) | |
Text
~
TENNESSEE VALLEY AUTHORITY CH ATTANOOGA, TENNESSEE 374o1 400 Chestnut Street Tower II d
August 10, 1981 Wd&'b' Director of Nuclear Reactor Regulation g]
AUG 171981 > 72 Attention:
Ms. E. Adensam, Chief g,
Licensing Branch No. 4 eouruces Division of Licensing ll 8
U.S. Nuclear Regulatory Commission Washington, DC 20555 m
Dear Ms. Adensam:
In the Matter of
)
Docket Nos. 50-327 Tennessee Valley Authority
)
50-328 As requested by NRC, TVA has reevaluated the issue concerning the potential for tornado generated missiles to compromise plant safety at Sequoyah Nuclear Plant by damaging the 480-V shutdown board transformers in the Auxiliary Building. My letter to you dated June 1, 1981 transmitted TVA's initial reevaluation of this lasue. We have since had time to care fully study the issue and have determined that some erroneous information was included. A revised writeup discussing our study is enclosed.
If you have any questions, please get in touch with D. L. Lambert at FTS 857-2581.
Very truly yours, TENNESSEE VALLEY AUTHORITY L. M. Mills, Manager Sworn to dsubsopibedbeforeme
~ this/O dayofdAM 1981 tTotary Public My Commission Expires Enclosure k[h(
8108180193 010810' PDR ADOCK 05000327 P
PDR An Equal Opportunity Employer
t ENCLOSURE SEQUOYAH NUCLEAR PLANT ASSESSMElff 0F THE POTENTIAL FOR DAMAGE TO THE 480-V SHUTDOWN BOARD TRANSFORMERS FROM TORNADO MISSILES The total event probability, P, of a vertical tornado missile impacting t
one of the intake and exhaust vents on the roof of the auxiliary building is the product of several conditional probabilities:
Pt=Pepapap a
n v
a where P is the probability of a tornado striking the plant 3
P is the probability of a missile striking a safety-related building given a tornado strike.
P is the probability of a vertical strike on a horizontal roof surface given a missile has impacted a safety-related building.
P is the probability of the missile striking that area of the roof containing a vent, given that a tornado missile has hit the roof.
Tornado Strike Probability (P,)
The torrado strike probability, P, can be estimated using the information 8
contained in References 1, 2, and 4.
The tornado strike probability for theSequogghNuclearPlant(SQN)areaisgiveninWASH-1300(Reference 1) i as 1 x 10 per year for any tornado intengity. This probability assumes an average tornt/o damage path area of 2.8 mi. This is the probability of occur-rence of tornadoes of any intensity. The fraction of tornadoes for given tornado tornadoes of any intensity. The fraction of tornadoes for given tornado intensities is given in Table 1.
Sophisticated methods of predicting sits specific tornado occurrence frequencies and frequencies for specific tornado intensities have been developed by-several_ authors. One such-nethod is incorporated in Reference 2 to predict the average occurrence frequencivs for tornadoes with intensities of F2 and larger.
Missile Strike Probability (P )
i The estimation of missile strike probabilities is based on the tornado missile risk analysis pr6sented in Reference 2.
Reference 2 considered a range of tornado intensities from F2 to F6 as having significant intensities (greater than 113 mi/h windspeed) to generate missiles.
A summary of the results of the analysis performed in Reference 2 is presented in Reference 3 An example case study is also presented in these references for a typical two-unit nuclear plant. The study assumes that unit 1 operates for the three years before completion of unit 2.
During this three-year construction period, the analysis aosumes 'a total of 5,000 available missiles from a spectrum of 26 missile types. Tais is more than l
}-
_1
4 the. number of missiles that would normally be available during the remaining life of the plant when both units are operating. For the typical auxiliary building in tornado Region I the pgobability of a tornado missile impact is given in Reference 2 as 1.56 x 10 per year during the three-year period when unit 2 is under construction. This missile strike probability includes the estimate of the tornado occurrence frequency for Region I.
Thus, the following estimate is reasonable for the Sequoyah auxiliary building.
-5 P
- P
= 1.56 x 10 per year 3
n The probability of a missile impact (P
- P ) for all targets of unit 1 in the example is given in Reference 2 asi Lower Limit = 4.g x 10 per year Mean = 8.42 x 10 per gar Upper Limit = 1.19 x 10 per year These upper and lower limits are the 95 percent confidence bounds for all targets.
Vertical Strike Probability (P )
y In order for a missile to pass through a roof opening it must first strike in the roof area of the building.
In addition, the missile trajectory must be nearly vertical in order for it to pass through a vent and pose a threat to the 480-V shutdown board transformers. The missile histories developed in Reference 2 show that a very small percent of all barrier impact events are on roof surfaces. From this, one can conservatively estimate the -
conditional probability of a vertical strike on the roof as P = 1 x 10 Probability of Striking a Vent (P )
l There is a total of 22 vents considered as potential missile entry points on the roof of the auxiliary building. Of these, eight have dimensions 4 feetx8feetand14gavedimensions4feetx4 feet. This is a total target area of 480 ft. The roof area of the auxiliary building in the 2
example problem is 44,000 ft.
It is reasonable to assume that a tornado I
missile has an equally likely probability of striking any area on the I
roof. The geometric probability of striking an area that contains a vent l
is then estimated as:
1 a
44 000 Applicability to SQN l
l It is recognized that the studies presented in References 2 and 3 are for a representative two-unit nuclear plant site and that the probability values l
presented are not developed specifically for SQN. However, the l
representative site layout used in Reference 2 does consider similar basic safety features as the Sequoyah plant with similar sizes and distances.
Also, the representative site is oriented with respect to the wind field in order to maximize potential missile trajectories. This is a conservatism due to the fact that not all tornado strikes at a specific site will be oriented in a similar manner. l
The studies in Reference 2 also estinate the probability of a missile strike for a representative one-unit plant. The results for this site configuration show only a 5 to 10 percent decrease over the values for a representative two-unit site. This demonstrates the strike probabilities are not very sensitive to variations in the specific plant target area.
Therefore, the calculation of site-specific probability values for SQN would not differ radically from those found in the representative site study.
The number and type of missiles available in also a site-specific consideration. As a conservative estimate of the tornado missile strike probabilities, the results for the case with one unit under construction are used.
Conclusion Using the probabilities discussed above, a conservative estimate of the probability of a vertical tornado generated missile striking a vent on the auxiliary building is:
Pt = (P
- P)*P
- P = (1.56 x 10-5) (0.1) (0.01) a n
v a
-0
= 1.56 x 10 per year The probability of the missile enterirg the opening at the exact trajectory to hit a safety-related component is thus extremely unlikely. Postulating a double strike event occurring in such a manner as to damage redundant transformers would certainly be less likely and is, therefore, not con-sidered to be a credible esent for SQN.
The event that a vertical missile strikes one of the vents on the auxiliary building roof does not guarantee that one of the 480-V shutdown board transformers will be damaged. Further conditional probabilities must be considered such as the deflection and ricochet of the missile after impacting a vent and the remaining energy available for causing damage to the transformers.
The concern regarding the presence of roof openings over the transformer rooms was recently reviewad for TVA's Nuclear Safety Review Staff. The review focused on the damage which could result from the passage of a single missile through one of the roof openings. Due to the presence of separation barriers between transformers and separation of electrical cables and conduits, no interaction of losses were found which could result in the loss of both trains of safety-related systems. The impact of a second missile in another transformer room was not considered credible for the reasons discussed above.
In summary, the probability of a tornado missile impacting a vent on the auxiliary building roof with the resulting damage leading to events which endanger the safe operation of the plant is believed to be negligibly small.
TABLE 1 WINDSPEED DISTRIBUTION Predicted Fraction Observed No. of of Total Fraction Observed Tornadoes for Region of Total Windspeed Classification (Reference 4)
I (Ref. 2)
(Ref. 4)
F6 (windspeed 318 ml/h) 0 0.001 0.0 F5 (windspeed 260 to 318 mi/h) 127 0.0012 0.005 F4 (207 to 260 mi/h) 673 0.0211 0.027 F3 (158 to 206 mi/h) 2,665 0.0713 0.107 F2 (113 to 157 mi/h) 7,102 0.2668 0.285 F1 (73 to 112 mi/h) 8,645 0.4405 0.346 F0 (40 to 72 mi/h) 5,718 0.1991 0.229. _.
. REFERENCES (1) WASH-1300, " Technical Basis for Interim Regional Tornado-Criteria."
(2) L. A. Twisdale, Tornado Missile Risk Study, EPRI NP-768, May 1978.
(3) L. A. Twisdale, W. L. Dunn, J.-Chu, " Tornado Simulation and R,'ak Analysis," ANS Topical Meeting on Probabilistic Analysis of Nuclear Reactor Safety, Los Angeles, California, May 1978.
(4) J. J. Tecsca, T. T. Fujita, and R. F. Abbey, Jr., " Statistics of U.S.
Tornadoes Based on the DAPPLE Tornaco Tape," American Meteorological Society, 11th Conference on Severe Local Storms, Kansas City, Missouri, October 1979.
i
}
_5_