ML17290A918

Justification for Revised Tornado Design Criteria,Rev 1.

ML17290A918
Person / Time
Site:	Columbia
Issue date:	12/27/1993
From:	Laframboise W, Noble L WASHINGTON PUBLIC POWER SUPPLY SYSTEM
To:
Shared Package
ML17290A891	List: ML17290A890 ML17290A918
References
NUDOCS 9401250140
Download: ML17290A918 (17)

Text

Justification for Revised Tornado Desi n Criteria Revision 1 pp,y, 4., g$ ~

W.L. LaFramboise I 4~>/Y>

Principal Civil Engineer Reviewed by g is/)3

.D. Noble S pervisor Civil Engineering Reviewed by F.D. Quinn S Meteorologist Approved by l "/>7/'l3

. Harness Manager Mechanical Systems 9401250i40 940i06' P.....DR,.

PDR ADOCK 05000397

L 4

LS

Justification for Revised Tornado Desi n Criteria 1.0 Back round The current tornado design criteria for WNP-2 is considerably more conservative than that required by the NRC or by probabilistic considerations of tornado hazards at the Hanford Site.

These conservatisms result in unnecessary costly designs to provide protection for the current design basis tornado. This document provides the basis for revising the tornado design criteria for new work or for re-evaluations of existing structures to more realistically reflect the tornado hazards at WNP-2.

The current NRC tornado design criteria applicable to WNP-2 are summarized in Section 2.0, the current WNP-2 criteria are summarized in Section 3.0, and the proposed revisions to the current WNP-2 criteria are presented in Section 4.0. The technical justifications for these revisions are provided in Section 5.0. The benefits of implementing these criteria revisions are summarized in Section 6.0.

2.0 NRC Tornado Desi n Criteria NRC tornado design criteria are defined in Regulatory Guide 1.76 [1] and Standard Review Plan (SRP) Sections 3.3.2 [2] and 3.5.1.4 [3]. These criteria have been established to provide an annual probability of exceedance of design loads of 1 x 10-7. To meet this objective, the NRC has defined in Regulatory Guide 1.76, three Tornado Intensity Regions which conservatively envelope the expected hazards within the respective regions. WNP-2 is in NRC Tornado Intensity Region III which has the least severe tornado loading requirements. A summary of NRC design requirements for Tornado Intensity Region lli are provided in the following sub-sections.

2.1 Tornado Wind and Differential Pressure Loads Regulatory Guide 1.76 specifies the following tornado wind and differential pressure loads for Tornado Intensity Region III:

loads resulting from a rotational wind velocity of 190 mph and a translational wind velocity of 50 mph for a maximum horizontal velocity of 240 mph a differential pressure load resulting from an external pressure drop of 1.5 psi at a rate of 0.6 psi /sec.

2.2 Tornado Missile Loads SRP Section 3.5.1.4 specifies that structures and equipment subjected to tornado effects are to be evaluated for resistance to damage from the following tornado generated missiles:

Impact Velocity for Tornado Intensity Region III Missile Wei ht Ibs Dimensions 4.3'orizontal ft/sec Wood Plank 115 3.6" x 0.94' 12'ong 190 6" Sch 40 Pipe 287 6.6" dia x 15'ong 33 1" Steel Rod 8.8 1" dia x 3'ong 26 Utility Pole 1124 13.5" dia. x 35'ong 85 12" Sch 40 Pipe 750 12.75" x 15'ong 23 Automobile 3990 16.4' 6.6' 134

These missiles are considered to strike surfaces in any direction. Vertical velocities are taken to be 70 percent of the horizontal velocities except the 1" steel rod which is assumed to have the same velocity in any direction. The utility pole and the automobile are considered to strike surfaces at any elevation up to a maximum elevation of 30 feet above the highest finished grade within 0.5 miles of the plant. All other missiles are considered to strike at any elevation above finished grade.

2.3 Load Combinations Based on SRP Section 3.3.2, total tornado loads resulting from wind loads, differential pressure loads, and tornado generated missiles are established from the following equations:

W Wp Wm Ww+ 0.5 W Ww+ Wm Ww + 0 5 Wp + Wm where:

W' Total tornado load Ww= Tornado wind load W> = Tornado differential pressure load Wm = Tornado-generated missile load 3.0 Current WNP-2 Tornado Desi n Criteria Current WNP-2 tornado design criteria are considerably more conservative than required by the applicable NRC Regulatory Guides and Standard Review Plans. Current tornado wind speeds are based on NRC Regulatory Guide 1.76 [1] Tornado Intensity Region I requirements (where as noted above, WNP-2 is in Tornado Intensity Region III). In addition, the load combinations used to define the net loads resulting from wind loading and differential pressure are conservative with respect to those required by NRC SRP 3.3.2 [2]. A summary of the current WNP-2 tornado design criteria is provided in the following sub-sections.

3.1 Tornado Wind and Differential Pressure Loads All Seismic Category I structures and equipment have been designed to withstand the effects of a design basis tornado defined by the following concurrent loads:

loads resulting from a rotational wind velocity of 300 mph and a constant translational wind velocity of 60 mph for a maximum horizontal velocity of 360 mph a differential pressure load resulting from an external pressure drop of 3 psi at a rate of 1.0 psi /sec.

The tornado wind velocities are considered to act horizontally and to be invariant with height. Wind forces are modeled in terms of static pressures based on ASCE paper No.

3269 [4].

3.2 Tornado Missile Loads All structures and equipment subjected to tornado effects are evaluated for resistance to damage from tornado generated missiles. Missiles which are considered consist of the following:

Horizontal Impact Missile Wei ht Ibs Dimensions Velocit ft/sec Utility Pole 1600 14" dia. x 35 ft long 241 1" Steel Rod 8 1" diax3ft long 259 These missiles are considered to strike surfaces in any direction. The utility pole is considered to strike surfaces at any level up to a maximum level of 30 feet above the highest finished grade within 0.5 miles of the plant. The steel rod is considered to strike at any level above finished grade.

An exception is made for the access door of the makeup water pump-house where the velocities of the above missiles are reduced to 85 ft/sec for the utility pole and 26 fl/sec for the steel rod. These velocities are consistent with those of Revision 2 of SRP Section 3.5.1.4 for Tornado Intensity Region III. These reduced velocities were incorporated to reduce the thickness of the door to facilitate equipment replacement while still providing adequate missile protection by virtue of the compliance to the SRP.

3.3 Load Combinations The total tornado load resulting from wind loads, differential pressure loads, and tornado generated missiles is established as follows:

W'= Ww NP = WP NP = W~

NP=W +Wp NP = W~+ Wm NP W~ + Wp + W~

where all terms are defined in Section 2.3.

4.0 Pro osed Revisions to the WNP-2 Tornado Desi n Criteria This section presents proposed revisions to the WNP-2 tornado design criteria for new work or re-evaluations of existing structures. These criteria are based on extensive studies which define site specific tornado hazards for the Hanford area. These criteria reduce unnecessary conservatism in the original WNP-2 design criteria and in the NRC criteria (which conservatively envelopes tornado hazards within Tornado Intensity Region III) while maintaining an annual probability of exceedance of design loads of 1 x 10-7. The technical justifications for these criteria are provided in the following section.

4.1 Tornado Wind and Differential Pressure Loads The following tornado wind and differential pressure loads will form the design basis for new work or re-evaluations of existing structures at WNP-2.

loads resulting from a rotational wind velocity of 157 mph and a translational wind velocity of 35 mph for a maximum horizontal velocity of 192 mph

a differential pressure load resulting from an external pressure drop of 0.70 psi at a rate of 0.24 psi /sec.

4.2 Tornado Missile Loads The following tornado generated missiles will form the design basis for new work or re-evaluations of existing structures at WNP-2:

Horizontal Impact Weight Velocity (ft/sec)

Missile Ibs Dimensions see discussion below Wood Plank 115 3.6" x 0.94' 12'ong 6" Sch 40 Pipe 6.6" dia x 15'ong 29.2 1" Steel Rod Utility Pole 287 8.8 1124 1" dia x 3'ong 4.3'50.7 13.5" dia. x 35'ong 27 33.1 12" Sch 40 Pipe 750 12.75" x 15'ong 23.8 Automobile 3990 16.4' 6.6' 95.9 These missiles are considered to strike surfaces in any direction. Vertical velocities are taken to be 70 percent of the horizontal velocities except the 1" steel rod which is assumed to have the same velocity in any direction. The utility pole and the automobile are considered to strike surfaces at any elevation up to a maximum elevation 30 above the highest finished grade within 0.5 miles of the plant. All other missiles are considered to strike at any elevation.

4.3 Load Combinations Load combinations for new work or evaluations of existing structures will be based on Standard Review Plan Section 3.3.2 [2] as summarized in Section 2.3. (Note that 1/2 of the tornado differential pressure load is combined with the tornado wind load as specified in the SRP.)

6.0 Technical Justification of Revisions in the Tornado Desi n Criteria WNP-2 is located in an extremely low tornado hazard area. The State of Washington experiences, on the average, less than one tornado each year. Within a one hundred mile radius of WNP-2, there have been only 21 reported tornadoes in the period from 1916 to 1980

[5]. Within the 570 square mile Hanford Site there has been only one reported tornado.

A study of the tornadoes in Washington, Oregon, and Idaho area was performed by Fujita [6]

which indicated that tornadoes occur primarily in "alleys". The location of these alleys with locations of reported tornadoes during the 20 year period from 1950 to 1969 is shown in Figure 5-1. Note that WNP-2 does not lie within any of the tornado alleys. These alleys are located in up-slope topography that we believe is necessary for tornado initiation within the Pasco Basin.

The Hanford Site is essentially in a down slope region from the Rattlesnake ridge making tornado initiation unlikely.

One of the first attempts at quantifying tornado hazards was made by Thorn [7]. By utilizing a simple geometric interpretation of probability, the probability of a tornado striking any given location is defined by the following equation:

P(s) =- an A

where:

P(s) = the probability that a tornado will strike a given location

a = the mean tornado damage path area in square miles n = thc mean number of tornadoes per year A = thc area of the local region in which thc mean number of tornadoes was reported in square miles By assuming independence between tornado wind speed and tornado strike probabilities, Markee, Beckerley, and Sanders expanded upon the Thorn model in Wash-1300 [8] to develop an interim regional tornado design criteria which currently forms the basis of Regulatory Guide 1.76. In their model, the probability of exceeding a given tornado wind speed is determined from the following equation:

P(v >V,) =P,(V,)P(s) where:

Pi(VJ = the probability of a tornado having wind speeds in excess of Vc. This probability is referred to as the tornado intensity distribution P(s) = the probability of a tornado strike as defined by the Thorn model Wash-1300 Methodolo The process employed in Wash-1300 to establish design tornado wind speeds for each Tornado Intensity Region is outlined below:

Tornado strike probabilities were determined for each five degree square of latitude and longitude in the contiguous United States based on a mean tornado damage path area of 2.82 square miles established by Thorn [7] for tornadoes occurring in a ten year period in Iowa. The mean number of tornadoes within each five-degree square was determined using 13 years of tornado occurrence records compiled by Pautz [9].

Based on data developed by NOAA climatologists (Ref [10] and [11]) for tornadoes which occurred within the contiguous United States during 1971 and 1972, tornado wind speed exceedance probabilities ( Pi(VJ ) were established independent of geographic location and tornado damage path area.

For each five degree square of latitude and longitude, a tornado wind speed was selected to provide an annual probability of exceedance of 1 x 10 7. Based on these wind speeds, design tornado wind speeds were established for each of the NRC's Tornado Intensity Regions.

Conservatisms in the Wash-1300 Methodolo The Wash-1300 tornado criteria is recognized as being conservative, and as is stated in Wash-1300, it was considered at the time of its development as an interim criteria until more realistic criteria could be developed. Major sources of conservatism of the Wash-1300 criteria as they relate to the Hanford area are as follows:

The mean path area for tornadoes occurring within the Pacific Northwest is considerably less than the 2.82 square mites assumed in Wash-1300. As noted above, this path area was based on tornadoes occurring in Iowa. As an example, in a study performed for the Jersey Nuclear Fuel Facility (now referred to as the Advanced Nuclear Fuels Facility) in Richland WA, Jaech [12] determined that the mean path area for tornadoes occurring within Washington, Oregon, and Idaho during 1950 to 1969 to be 0.48 square miles.

Tornado wind speeds within the Columbia Basin are considerably lower than the national average. For example, based on the data reported in [6] for the 20 year period from 1950 to 1969, within the region east of the Cascade Mountains and North of the 45+ parallel, the highest reported tornado intensity was F2 (wind speeds113-157 mph). During this period, there were only fourreported F3 tornadoes (wind speeds 158- 206) within Washington, Oregon and Idaho. Two of these tornadoes occurred in the Portland/ Vancouver area, one occurred in the Seattle/ Tacoma area, and the other occurred at the Oregon/ Idaho border near Ontario, Oregon.

The Wash-1300 methodology assumed that the maximum wind speed of a tornado occurs uniformly over the entire damage area. However, wind speeds vary across, and along the path of a tornado. For example, based on a simple Rankine vortex model of tornado winds and data obtained from the Super Outbreak of Tornadoes of April 3-5, 1974 [13], McDonald [14] established that for a F3 tornado, only 17 percent of the damage area experiences, winds that are in F3 range.

5.5 Refined Tornado Hazard Models Substantial refinements have been made by Abbey and Fujita [15] and by McDonald

[14] in the methodology for predicting tornado hazards since the time the Wash-1300 study was performed. The refinements incorporated in these hazard models are summarized below:

a larger database of reported tornadoes is available for statistical analyses than was available at the time the Wash-1300 study was performed estimates of the number of unreported tornadoes are made and are included in ~

the hazard models based on population density and land use regionalized path area - intensity and occurrence frequency - intensity relationships are developed directly from reported data and estimated unreported tornadoes the variation of wind speed across and along the tornado path are explicitly accounted for in the revised models The Department of Energy (DOE) through Lawrence Livermore National Laboratory contracted Fujita and McDonald to perform independent tornado hazard evaluations for the various DOE nuclear related sites using their tornado hazard models. The results of this effort were published in 1985 [16]. As a result of the low tornado hazard for the Hanford area, the maximum horizontal wind speeds derived independently by Fujita and McDonald were 179 mph and 177 mph respectively for an annual probability of exceedance of 1 x 10-7.

In addition, the American National Standards Institute has proposed a "Standard for Tornado and Extreme Wind Characteristics at Nuclear Power Plant Sites" [24] which estimates tornado wind speeds of 180 mph corresponding to an annual probability of exceedance of 1 x 10-7 for the region west of the Rocky Mountains.

Pacific Northwest Laborato Tornado Climatolo Stud Pacific Northwest Laboratory (PNL) [22] compiled data on the characteristics of tornadoes occurring in the contiguous United States for the period from January 1, 1954, through December 31, 1983 based on reported tornadoes contained in the National Severe Storms Forecast Center tornado data base. Statistical analysis of, these data were performed to estimate tornado wind speeds for each 5 degree square of latitude and longitude with 10 s, 10~, and 10-" probabilities of exceedance per year both mean and 90 percentile confidence limits. Since design wind speeds have

'or traditionally been based on a mean probability of exceedance of 10-" (i.e., WASH-1300

4 A,

h

[8], McDonald [14], and the Department of Energy [16]), the results presented below are based on this value.

The statistical models used in the PNL analysis were similar to those used in the WASH-1300 study with the following exceptions:

rather than assuming a mean tornado path area of 2.82 square miles, the PNL study calculated expected mean tornado path areas in each 5 degree squareby fitting a log-normal probability distribution to the recorded tornadoes within the square and integrating over the distribution to establish the expected mean path area. The computed expected mean values are generally substantially smaller than 2.82 square miles.

in Wash-1300, tornado wind speed exceedance probabilities ( Pi(V) ) are established for the entire contiguous United States from the relative frequencies of each tornado intensity class (i.e., F-scale) and are assumed to be in the form of a log-normal probability distribution. In the PNL study, tornado wind speed exceedance probabilities are more realistically based on relative frequencies of the total affected area of each tornado intensity class and are based on a Weibull probability distribution. In addition, the PNL study develops independent tornado intensity distributions of western and eastern regions of the contiguous United States.

The PNL studies result in significantly lower tornado wind speeds than that predicted by WASH-1300. Tornado wind speeds for a probability of exceedance of 10 > for the 5 degree square centered at 47.5O North, 122.5O West is 186 mph and for the 5 degree square centered at 47.5'orth, 117.5O West is 192 mph (note that WNP-2 is located in the latter square adjacent to the border of these two squares). The slightly higher wind speeds developed by PNL than by Fujita and McDonald can be attributed to the differences in the methodology and the following conservatisms in the PNL approach:

the PNL approach does not account for the variation of wind speed across and along the tornado path. As noted above, an F3 tornado will have only 17 percent of the damage area experiencing winds in the F3 range. It is estimated that approximately 36 percent of the path length and 48 percent of the path width will be in the F3 range. In the PNL study, it is reasoned that the conservatism of neglecting variations along the path length will offset potential miss-classification errors of violent tornadoes (i.e., reporting tornadoes with F-scales of 4 and 5 as having a lower intensity) as reported by Twisdale et al.

[23]. This compensation should be overly conservative west of the Rocky Mountains since there have been only rare reports of violent tornadoes in this region. (Fujita [17] reports only 3 violent tornadoes in the period form 1916 to 1977 all of which were located in the Imperial Valley of southern California. No violent tornadoes were reported west of 105'est in the period from 1953 to 1983 [22]). In any case, neglecting variations across the path width as is performed in the PNL study will result ln an over estimate of areas effected by peak tornado winds.

expected tornado path areas are calculated by fitting a log-normal probability distribution to the recorded data and integrating over the distribution to establish the expected mean path area. This approach would provide a more accurate estimate of the mean than a simple arithmetic average provided that the probability distribution accurately models the data. As noted in the PNL report, the accuracy of a log-normal distribution is open to question and a Weibull distribution may be a more accurate model. Based on data contained in Appendix A of the PNL report, the use of a log-normal distribution potentially results in a over prediction of the expected area by a factor of 1.5 to 2.0. For

the Pacific Northwest, this would result in an over prediction of wind speeds by approximately 10 mph.

A iicabili of Refined Tornado Hazards Studies to WNP-2 Each of the above refined tornado hazards studies are directly applicable to the geographic region containing WNP-2. Although these studies were independently performed using different probabilistic models, they are in close agreement with respect to the prediction of tornado hazards for the WNP-2 site. The highest predicted wind speed for 1 x 10 7 annual exceedance probability is 192 mph while the lowest is 177 mph. These values represents approximately an 8 percent difference from the highest value to the lowest value. As discussed above, the 192 mph wind speed developed by PNL is slightly more conservative than those predicted by Fujita and McDonald due to the conservatism in their model. The proposed design tornado wind speed for WNP-2 of 192 mph was selected since it is the most conservative of the above predictions.

The use of this wind speed as the WNP-2 design basis will conservatively meet the NRC objective of maintaining a 1 x 10-7 annual exceedance probability of tornado wind speeds.

5.2 Tornado Wind and Differential Pressure Desi n Parameters Tornado wind and differential pressure design parameters can be established as a function of the maximum horizontal wind speed. For this purpose, the Rankine Vortex

[14] and [19] tornado wind field model will be used. In this model, the translational and rotational wind velocities are determined from the following equations [26]:

VT +V=V v=(v'v,'}'"

1 VR =-VH8 2

where:

Vro = Rotational Wind Velocity Vmax = Maximum Horizontal Wind Velocity VT = Translational Wind Velocity VR = Radial Wind Velocity Vo = Tangential Wind Velocity Utilizing these equations and a maximum horizontal wind speed of 192 mph, the resulting rotational wind velocity is 157 mph and the resulting translational wind velocity is 35 mph.

The maximum differential pressure is established based on the tangential component of the wind velocity from the following equation:

hp =pVy where

p = mass density of air For a maximum rotational wind velocity of 157 mph, the above equations result in a maximum differential pressure of 0.70 psi [26]. The rate of pressure change is established from the following equation:

dp, Clt

=p V~

VT Rp where Ro = radius of maximum winds Based on Regulatory Guide 1.76 [1], R~ is assumed to be 150 ft. Utilizing this value, the rate of pressure change is 0.24 psi/sec [26]. Note that these values appear to be very conservative considering the almost total lack of damage from northwestern tornadoes.

5.3 Tornado Missiles The spectrum of proposed design basis missiles is identical to those contained in Standard Review Plan 3.5.1.4 except that they are applicable to a maximum wind speed of 192 mph. The proposed missile velocities are based on analyses performed by the Supply System [26]. These analyses utilized an identical methodology to that used by the National Bureau of Standards [19] (NBS) to develop the tornado missile velocities contained in Standard Review Plan Section 3.5.1.4 for tornadoes with maximum wind velocities of 240, 300, and 360 mph respectively for each of the three NRC tornado intensity regions. The NBS methodology was implemented in Supply System computer program TMISSILE. This program was verified by comparing predicted missile velocities to those obtained in [19] for the tornado wind speeds evaluated in that document and to closed form solutions of simple aerodynamic problems.

The conservatism of the NBS methodology is verified by the extensive research sponsored by EPRI ([20] and [21]) to establish more realistic tornado missile criteria.

Major improvements in the EPRI tornado missile analysis methodology are outlined below:

The EPRI tornado wind field model more realistically considers near ground tornado winds and models variations of wind velocities as a function of height. The NBS evaluations assume a simple Rankine vortex model of tornado winds which defines wind speeds strictly as function of radius from the center of the tornado. In the NBS evaluations, the missiles are conservatively assumed to be introduced into the wind field at an initial elevation of 40 meters. As a result, during their free fall, the missiles are subject to tornado winds which are invariant with height. Missile velocities were also established by EPRI for missiles with an initial height of 40 meters. However, using the more refined EPRI wind field model, the forcing function applied to the missiles considers expected variations with height, which reduces in magnitude as the missile approaches the ground.

The NBS flight model assumes that the missile is acted on strictly by drag forces.

Lift and tumbling effects are neglected. Drag - Area coefficients are assumed to be an average of the analytically derived coefficients for three orthogonal directions of the missile. In the EPRI flight model, all 6 degrees of freedom of missile flight are 10

explicitly included in the model. Flight parameters and drag coefficients were experimentally determined from wind tunnel testing.

Tornado missile velocities calculated using the EPRI wind field and missile flight model consistently bound those obtained using the NBS methodology. This result substantiates the inherent conservatism of the NBS methodology.

Since the Supply System utilized a methodology identical to that used in the development of Standard Review Plan Section 3.5.1.4 and this methodology is generally regarded as being conservative, the proposed design basis missiles speeds provided a conservative definition of the missile hazards at WNP-2.

6.4 Load Combinations The load combinations specified in the revised criteria are identical to those in SRP Section 3.3.2. These load combinations recognize the fact that peak wind and pressure differential loads do not occur simultaneously as was conservatively assumed in the original WNP-2 criteria.

6.0 Benefits of Im lementin the Revised Tornado Desi n Criteria By implementing the proposed revisions, the tornado design criteria will more realistically reflect the tornado hazards at WNP-2 while maintaining the NRC's objective of providing an annual probability of exceedance of design loads of 1 x 10-7. These revisions will result in more realistic wind velocities, differential pressure loads, missile velocities, and load combinations.

The proposed revisions will benefit the design of exterior structures and components that must consider postulated tornado effects. Significant benefit is expected during the implementation of future plant modifications where relief of tornado design criteria will facilitate construction. As an example, the temporary relief from tornado design criteria granted for the installation of the diesel fuel oil polishing system facilitated the installation of this system and signiTicantly reduced the cost and time required for the installation [25].

7.0 References U.S. Nuclear Regulatory Commission Regulatory Guide 1.76, "Design Basis Tornado for Nuclear Power Plants,", Washington D.C., April 1974

[2] U.S. Nuclear Regulatory Commission Standard Review Plan Section 3.3.2, "Tornado Loadings," NUREG-0800, Rev 2Washington D.C., July 1981

[3] U.S. Nuclear Regulatory Commission Standard Review Plan Section 3.5.1.4, "Missiles Generated by Natural Phenomena," NUREG-0800, Rev 2, July 1981, Washington D.C.

[4] ASCE Paper No. 3269, "Wind Forces on Structures," Transactions of the American Society of Civil Engineers, Vol. 126, Part II, 1961.

Stone, W.A., Thorp, J.M., Gifford, O.P., Hoitink D.J., "Climatological Summary for the Hanford Area", Prepared for the U.S. Department of Energy under Contract DE-AC06-76RLO 1830 by Pacific Northwest Laboratory Operated by Battelle Memorial Institute, PNL-4622- UC-11.

[6] Fujita, T.T., "Estimate of Maximum Wind Speeds of Tornadoes in Three Northwestern States",

SMRP Research Paper No. 92, The University of Chicago, December 1970.

[7] Thorn, H.C.S., "Tornado Probabilities," Monthly Weather Review, Vol. 91, pp 730-736, 1963.

[8] Markee, E.H., Beckerley, J.G., and Sanders, K.E., "Technical Basis for Interim Regional Tornado Criteria," WASH-1300, U.S. Atomic Energy Commission, Office of Regulation, Washington, D.C., 1974.

[9] Pautz, M.E., ed., "Severe Local Storm Occurrences 1955-1967," ESSA Technical Memo, WBTM FCST 12, Office of Meteorological Operations, Silver Springs, MD.

[10] Fujita, T.T., "F-Scale Classification of 1971 Tornadoes," SMRP Research Paper No. 100, 1972.

[11] Fujita, T.T., Pearson, A. D., "Results of FPP Classification of 1971 and 1972 Tornadoes," Eight Conference on Severe Local Storms, Denver, CO, 1973.

[12] Jaech, J.L., "Statistical Analysis of Tornado Data for Three Northwestern States," Jersey Nuclear Co., 1970.

[13] Fujita, T.T., "Super Outbreak'Tornadoes of April 3-4, 1974," Department of the Geophysicai Sciences, The University of Chicago, Chicago, IL, 1975.

[14] McDonald, J.R., "A Methodology for Tornado Hazard Probability Assessment," Institute for Disaster Research, Texas Tech University, Prepared for U.S. Nuclear Regulatory Commission, NUREG/CR-3058, 1983. Ih

[15] Abbey, R.F., Fujita, T.T., "The DAPPLE Method for Computing Tornado Hazard Probabilities:

Refinements and Theoretical Considerations," Preprints of the Eleventh Conference on Severe Local Storms, American Meteorological Society, Boston, MA, 1979.

[16] Coats, D. W., Murray, R. C., "Natural Phenomena Hazards Modeling Project: Extreme Wind/Tornado Hazard Models for Department of Energy Sites," UCRL-53526, Rev 1, Lawrence Livermore National Laboratory, Livermore CA, August 1985.

[17] Fujita, T,T. "Workshop of Tornadoes and High Winds for Engineering Applications," SMRP No.

165, Department of the Geophysical Sciences, The University of Chicago, Chicago, IL, 1978.

[18] McDonald, J.R., "Flight Characteristics of Tornado Generated Missiles," Institute for Disaster Research, Texas Tech University, Lubbock, Texas, 1975.

[19] Simiu E., and Cordes, M., "Tornado-Bom Missile Speeds", NNBSIR 76-1050, Institute for Basic Standards, National Bureau of Standards, Washington D.C., April 1976.

[20] "Wind Field and Trajectory Models for Tornado-Propelled Objects," EPRI Report EPRI NP-748, Electric Power Research Institute, Palo Alto, CA, May 1978.

[21] "Wind Field and Trajectory Models for Tornado-Propelled Objects," EPRI Report EPRI NP-2898, Electric Power Research Institute, Palo Alto, CA, March 1983.

[22] Ramsdell, J.V., Andrews, G.L., "Tornado Climatology of the Contiguous United States", Pacific Northwest Laboratory Operated by Battelle Memorial Institute, Prepared for U.S. Nuclear Regulatory Commission, NUREG/CR-4461, 1986.

[23] Twisdale, L.A., Dunn W.L., Lew, S.T., Davis, T.L., Hsu, J.C., and Lee, S.T., "Tornado Missile Risk Analysis Appendices", Electric Power Research Institute, Palo Alto California, EPRI-769.

[24] "Standard for Estimating Tornado and Extreme Wind Characteristics at Nuclear Power Sites,"

ANSI/ANS-2.3-1983, American Nuclear Society, La Grange Park, illinois.

[25] Supply System Letter to the U.S. Nuclear Regulatory Commission, "WNP-2 Operating License NPF-21 Request for Temporary Relief from Tornado Missile Design Criteria," G02-93-028, February 9, 1993.

[26] Supply System Calculation "Plant Specific Tornado Missile Speeds for a Design Basis Tornado with a Maximum Velocity of 192 mph," Calculation CE-02-93-28, Rev 0.

13

<<i FIGURE 6-1 LOCATION OF TORNADO ALLEYS WITH REPORTED TORNADOES FROM 1950 TO 1969 Reference WNP-2 FSAR It@ It4 Ito I I4 I IO SD 50 II 1

46 V 46 tkNC WaSIIINSTON "Ir 4hgstIaev <<MICOVO I

I 0 Ni I I

r' I ls 10 kNO pj j II I'<<I 44 nl OREGON

~

I' r

)

V

/ r V

44 V

VV rv v VV

,i g3 f"o " v I

IV V ~

V I

I I

Y Y I 4t 4t V J.

Ito I IO I ICs I I4 loo toO Ultto 4<<aaacvcaitaT.OW O~OO4OOCS

~ Ot<<

<<~San r4, ~0 ~4 r+

tv<<<<CI 4 OOI 4 as 4 ' '

STORMS IN VCAW %OX ARC DttltaCD m ToaWuaSS iN TRIS SNSST W VVATSR OOOIIT TOICNADOCS ORC IN SOVARC MlSCS F SCJI1 C VIIND SOCCO IN M.O.M,