tornado

Wind/Tornado Design Criteria Development to Achieve Required Probabilistic Performance Goals

DOROTHY S. N G

Lawrence Livermore National Laborator),, Livertnore, California 94550

1. INTRODUCTION

The U.S. Department of Energy (DOE) has established specific requirements for the design of structures, systems, and components (SSCs) of critical facilities for resisting all types of internal and external natural hazard events. The Lawrence Livermore National Laboratory (LLNL) was chartered to develop design guidelines for natural hazards utilizing the best technical knowledge and engineering judgment to assure that the design of the SSCs conform to performance goals set forth by DOE.

This report documents the strategy employed to develop wind/tornado hazard design criteria for a critical facility to withstand loading induced by the wind/tornado hazard. These guidelines were developed by a Wind/Tornado Work- ing Group (WTWG) at LLNL composed of six experts who are knowledgeable in the fields of structural engineering, wind/tornado engineering, and meteorology. Utilizing their best technical knowledge and judgment in the wind/tornado field, they met and discussed the methodologies and re- viewed available data. A review of the available wind/ tornado hazard model for the site, structural response evaluation methods, and conservative acceptance criteria led to a proposed design criteria that has a high probability of achieving the required performance goals.

2. DESIGN REQUIREMENTS

The primary considerations of the design of the critical facility are public safety, worker safety, and environmental protection. The DOE design requirements document serves as the basis for design of the facilities. The contractors may propose methods which offer improvements over the speci- fied requirements. However, if such methods are proposed, the contractor is required to demonstrate the validity of resulting improvements.

The Tornado: Its Structure, Dynamics, Prediction, and Hazards. Geophysical Monograph 79 Copyright 1993 by the American Geophysical Union.

The pertinent wind/tornado design requirements are sum- marized below:

1. The design of facilities shall encompass annual safety performance goals of (1) 5 x 10 -7/yr probability for an early fatality to an average individual assumed to be located within 1 mile (--- 1.6 km) of the reactor facility control perimeter; (2) 2 x 10-6/yr probability for a long-term fatality to plant workers or a member of the general public located within 10 miles (---16 km) of the reactor facility, but outside of the control perimeter; (3) 1 x 10-6/yr probability for a large release of radioactive materials from a reactor accident; and (4) 1 x 10-5/yr probability for core damage.

2. DOE encourages the contractors to incorporate the knowledge and experience over many years in nuclear reactor design accumulated.

3. DOE requires the contractors to make use of past regulatory compliance in commercial reactor design or demonstrate improvement over regulatory provisions.

The design requirements document stated that external events shall be addressed proactively in the design process in the context of the large-release goal. To investigate the achievement of the large-release goal requires additional consideration of containment failure modes, namely, direct, bypass, and penetration failures. It is more realistic to show the achievement of the core damage goal as an interim step, by wind/tornado risk assessment. The WTWG expects that the mean large-release risk frequency will be at least an order of magnitude lower than the mean core damage risk frequency because it is unlikely that wind and missile effects will significantly contribute to breaching containment. Therefore achieving the core damage goal is believed to ensure achievement of the large-release goal.

3. DESIGN GUIDEI.INES DEVELOPMENT STRATEGY

To ensure the safety of the plants, DOE requires that facility design meet stringent safety goals and should incorporate reactor experience and compliance with regulatory requirements in facility design. An effective approach to satisfy these requirements is to employ a panel of experts

400 WIND/TORNADO DESIGN CRITERIA

who have technical knowledge, experience, and judgment as a result of years of research and service in this area. Thus the WTWG was established for the development of wind/ tornado design guidelines. The WTWG consists of a group of six nationally recognized technical experts; they are briefly described below in alphabetical order:

Robert F. Abbey, Jr., is a well-known meteorologist who is currently serving as the director of Marine Meteorology Research, Office of Naval Research. He specializes in probabilistic and statistical analyses of the occurrence of natural phenomena.

W. Lynn Beason is associate professor in the Civil Engi- neering Department of Texas A&M University. His expertise is in wind-borne missile research, tornado risk analysis, and structural design.

T. Theodore Fujita is a world-known tornado assessment expert. Fujita developed the Fujita scale (F scale) for estimating the relative intensity of tornadoes and the DAPPLE method for estimating the probability of tornado occurrence.

Dale C. Perry, chairman of the WTWG, is the head of the Department of Construction Science of Texas A&M Univer- sity. His expertise is in structural design and wind engineering. He has over 20 years of experience in postdisaster investigations.

John W. Reed is an associate at Jack R. Benjamin & Associates, Inc., California. He has 25 years of experience in nuclear reactor design and probabilistic risk assessment (PRA) for seismic and wind hazards. Reed performed the preliminary wind speed PRA for this project.

Lawrence A. Twisdale, Jr., is the senior vice president of Applied Research Associates, Inc., Southeast Division in North Carolina. He specializes in tornado missile simulation and design methodology for tornado hazard assessment. Twisdale performed the preliminary missile criteria PRA for this project.

Additionally, James McDonald, a professor and re- searcher at Texas Tech University, is a member of the senior review group for this project; he provides guidance to the WTWG and reviews reports produced by the group. McDon- ald specializes in wind hazard assessment and missile impact research.

Shortly after formation of the WTWG and prior to their first meeting, the existing wind/tornado hazard assessment reports, Nuclear Regulatory Commission (NRC) provisions, design standards, and other references available for the site of interest were distributed to the WTWG. The methodologies and data utilized in existing regional and site-specific hazard assessments were compared and discussed by the WTWG in a series of meetings. Then, using their cumulative technical knowledge and judgment in the wind/tornado field, the WTWG planned the development of wind/tornado design guidelines.

The six basic goals of the WTWG strategy are as follows: (1) to build in conservatism by reducing wind/tornado goals to 1/10 of total project goals; (2) to specify design wind load specifications based on site-specific hazard frequency curves; (3) to specify design missile load specifications based

on DOE guidelines and the Electric Power Research Insti- tute (EPRI) NP-2005 report; (4) to define conservative structural response evaluation and acceptance criteria to substan- tiate the design guidelines; (5) to assure the conformance of design guidelines with safety goals by performing preliminary probabilistic risk assessments; and (6) to make use of nuclear industry standards, NRC provisions, and national codes and standards as benchmarks and references.

3.1. Developing Wind Load Specifications Based on Site-Specific Hazard Frequency Curves

In the 1970s the NRC divided the U.S. into three geographical regions and recommended design basis tornado (DBT) wind speeds for each region in NRC Regulatory Guide 1.76 [U.S. Nuclear Regulatory Commission, 1974]. These DBT wind speeds are believed to be quite conservative, resulting from coarse assumptions and problems with the regional data sets employed.

More recently, the Office of Nuclear Reactor Regulation [1988] position paper divided the U.S. into four regions and recommended DBT wind speeds based on regional analyses. Recommendations based on regional analyses cannot pro- vide a realistic portrayal of the risk associated with specific sites within the region.

In 1979, two leading wind engineering experts, Fujita and McDonald, performed independent site-specific wind/ tornado hazard assessments for 26 DOE facilities. The site of interest was among them. These analyses utilized advanced methodologies. Two of the most important improvements involved the use of wind speed gradation, both across the width and along the length of tornado damage paths. In addition, techniques were introduced to allow adjustments to be made to tornado data sets to account for unreported tornado events. Based on these hazard assessment results, Kennedy et al. [1989], in their DOE-UCRL-15910 report, provided design and evaluation guidelines for hazard frequencies up to 2 x 10-5/yr.

In 1985, a third site-specific hazard assessment was performed by Twisdale for the site of interest. This analysis used computerized simulation techniques to estimate hazard probabilities and accounted for building size effects.

In summary, wind assessment technology has evolved from overly conservative regional analyses to realistic site- specific assessments. In this process, coarse assumptions have been refined by application of accumulated knowledge appropriate to tornado hazard assessment. In addition, during the ensuing years since the introduction of NRC Regu- latory Guide 1.76, more complete tornado occurrence data have been recorded, allowing more accurate assessments. The resulting site-specific analyses, which take these advancements into account, are improvements over earlier regional analyses. Therefore the WTWG opted to incorporate site-specific hazard curves in the design guideline development.

The use of site-specific hazard frequency curves in design guidelines development reflects the accumulated

NG 401

and experience in wind engineering. The recommended design guidance presented in this paper is a major improvement to NRC provisions. The conformance of these proposed design guidelines for design was confirmed by preliminary PRAs described later in this paper.

It must be noted in wind hazard assessment that limita- tions of tornado intensity and occurrence data, F scale wind speed assumptions, and analysis processes incorporated in the hazard models result in assessment uncertainties.

3.2. Developing Missile Load Specifications Based on DOE Guidelines and EPRI Tornado Missile Simulation Report

The Offie of Nuclear Reactor Regulation [1981] Standard Review Plan (SRP), NUREG-0800, recommends a conservative spectrum of tornado-generated missiles and corresponding velocity fractions. There is no regional distinction in the SRP missile types and sizes. However, the missile impact velocities vary on the basis of regional DBT wind speeds. DOE-UCRL-15910 [Kennedy et al., 1989] also recommended missile spectra for various categories of facilities based in part on results of missile research performed at Texas Tech University. The missile spectra were developed for risk frequencies up to 2 x 10-5/yr.

The Electric Power Research Institute sponsored development of a probabilistic tornado missile risk methodology to assess potential missile hazard for nuclear power plants. Results of this effort are presented in the EPRI NP-2005 report [Twisdale and Dunn, 1981]. Missile impact velocities were estimated on the basis of computer simulations of potential missile populations at nuclear power plants and computer-generated missile trajectories. The velocity fractions for steel pipes generated by EPRI are slightly larger than SRP velocity fractions.

In summary, state-of-the-art missile simulation technology incorporating relevant data bases and newly developed experimental results is an improvement over the technology used for NRC missile impact estimates in the 1970s. In the judgment of the WTWG, the recommended missile spectrum and the maximum impact velocities guidelines should be formulated by modification of DOE UCRL-15910 guidelines. The modification consists of incorporation of additional missiles and extrapolation from the EPRI tornado missile simulation results.

3.3. Assuring Conformance of Design Guidelines by Performing Preliminary Probabilistic Risk Assessment

Probabilistic risk assessments were performed to estimate annual damage risk frequencies for the facilities based on the recommended design guidelines. These wind damage risk frequencies are measures of achievement of the performance goals. To ensure the conformance of the design with the required performance goals, independent preliminary PRAs were performed to evaluate both wind speed and missile impact.

4. DESIGN GUIDELINES TECHNICAL SUPPORTING BASES

The objective of the design guidelines development is to recommend guidance for the design of the facility to meet the safety goals set forth by DOE. The basic philosophy for design guidelines development is to select low wind load specifications. The design goal is met by applying conservative response evaluation and acceptance criteria.

Core damage guidelines were developed for the structures, systems, and components required to bring a plant to a safe shutdown and to maintain this condition following a wind or tornado event. Production guidelines were developed for SSCs which cannot be cost effectively repaired or replaced within 90 days following a wind or tornado event. Ordinary SSCs, which are classified as low-hazard SSCs, must satisfy only UCRL-15910 wind/tornado design guidelines.

4.1. Wind Load Specification At the work group meetings the WTWG held long discus-

sions to establish the proper direction of wind/tornado design guidelines. Based upon references provided to the working group and the combined judgment of the hazard assessment and the risk assessment analysts, the following major factors were taken into account: (1) conservatism embedded in methodologies of assessment, (2) uncertainties identified in the assessment, (3) site-specific wind speeds in the geographical areas of interest, (4) site-specific wind speeds which achieve risk goals, and (5) acceptable wind speeds by the consensus of WTWG members.

After careful study of the hazard frequency curves and the examination of preliminary risk calculations, through an iterative process, the WTWG selected a tornado fastest-mile wind speed of !80 miles per hour (mph) (1 mi. = ---1.6 km), which corresponds to a probability of occurrence of 3 x 106/yr, and it achieves the core damage goal of 1 x 10-6/yr.

A preliminary risk analysis on the production event indi- cates that straight wind controls the design. Using the straight wind model, a fastest-mile straight wind speed of 130 mph was selected. This wind speed corresponds to an occurrence probability of 2 x 10-4/yr.

4.2. Missile Load Specification

After evaluating the SRP provisions, EPRI missile simulation results, and DOE-UCRL-15910 missile spectra, the WTWG formulated missile guidelines based on the following supporting bases: (1) selection of missile types that have been observed in past tornado damage investigations, (2) modification of the DOE-UCRL-15910 to extend the risk frequency from 2 x 10-5/yr to 1 x 10-6/yr, and (3) extrapolation of the EPRI tornado missile simulation results.

As a result of this effort, the WTWG added an 8-in.- diameter steel pipe to the DOE UCRL-15910 missile spectrum to extend the risk frequency to 1 x 10-6/yr. The horizontal impact velocity for the 8-in. steel pipe was ob- tained by extrapolating EPRI results for a

402 WIND/TORNADO DESIGN CRITERIA

TABLE 1. Wind-Borne Missiles Criteria

Missile Missile Criteria Safety Production

2 in. x 4 in. timber plank 12 ft long, 15 lb

3-in.-diameter standard steel pipe, 10 ft long, 75 lb

8-in.-diameter standard steel pipe, 15 ft long, 430 lb

Automobile, 3000 lb

horizontal impact speed, mph 150 effective height, ft 200 vertical impact speed, mph 100 horizontal impact speed, mph 120 effective height, ft 150 vertical impact speed, mph 80 horizontal impact speed, mph 100 effective height, ft 75 vertical impact speed, mph 70 rolls/tumbles along ground, mph 35 effective height, ft 30 vertical impact speed, mph 20

115 200

75

90 150 60

75 30 50

25 30 15

Metric conversions: 1 in. = ---2.5 cm; 1 ft = ---0.3 m; 1 lb = ---0.45 kg. Horizontal and vertical missile speeds are uncoupled and should not be combined. Specific missiles

represent a class of debris. Effective heights are elevations above ground level.

steel pipe at the 90 percentile. This extrapolation included conversion from a maximum velocity of 300 mph to 200 mph. The vertical impact velocity was set at 2/3 of the horizontal impact velocity by the WTWG.

Table I presents spectra of potential missiles with recommended impact velocities and corresponding effective heights (metric conversions are provided in a footnote). The barrier thickness shall be determined by equations given by Twisdale and Dunn [1981]. The minimum thickness of reinforced concrete barriers is 12 in. for all elevations below 75 feet, and 8 in. for all elevations above 75 feet. The minimum thickness of steel barriers is 1/2 in. for all elevations.

4.3. Structural Response Evaluation and Acceptance Levels

The WTWG incorporated conservative structural response evaluation methods and acceptance levels as follows:

1. Account for the fact that the critical facility is a high-hazard facility by increasing the importance factor to 1.35.

2. Use exposure C for tornado-controlled designs by assuming that the facilities are located in open terrain with obstruction having heights generally less than 30 feet. This assumption accounts for uncertainties in the characteristics of tornadoes.

3. Account for the variability in structural resistance properties by using a load factor of 1.3 on design wind pressures.

4. Select the velocity pressure exposure coefficient and the gust factor corresponding to the height of structures for both tornado and straight wind controlled designs.

5. Account for the uncertainty in missile impacts by following the American Concrete Institute (ACI) 349-85 recommended 20% increase in calculated barrier thickness.

6. Since a conservative missile spectrum and maximum tornado-generated impact velocities were selected for bar-

rier design, the load factor of 1.3 does not apply to missile loads.

The SSCs shall be designed for the most severe total wind load (W t) weight. This is a combination of wind pressure (W,,), atmospheric pressure change (APC) (Wp), and missile impact loads (Win). The load factor applies only to wind pressure Ww and APC pressure W,, but not to missile loads. In addition, the APC pressure load is half of its maximum value at the radius of maximum wind speed in a tornado; therefore, only 0.5W, is combined with wind pressure and/or missile impact, as given below:

Wt = 1.3 W,,,

Wt = 1.3 Wp

Wt =Wm

Wt = 1.3W,, +Wm

Wt = 1.3 W,,, + 0.5 x 1.3 Wp

Wt = 1.3Ww + 0.5 x 1.3W, +Wm In the case of events controlled by straight wind, the APC pressure load does not exist, and the load combinations are as follows:

Wt = 1.3 Ww

Wt -- Wm

Wt = 1.3 Ww + Wm

4.4. Risk Assessment for Wind Speeds and Missile Criteria

The preliminary PRAs were performed during the concep- tual design stage; hence, component and plant details

NG 403

not available. This limits the applicability of the preliminary PRAs to the final design. In addition, the results must be considered to be preliminary owing to the following simplifications which were built into the wind speed risk analyses:

1. The preliminary PRA analysis was simplified by using the results from previous PRAs to obtain an approximate but useful estimate of the expected damage frequency.

2. The preliminary mean damage frequency, which was used, represents damage within the 60-90% probability range.

3. Mean damage fragility curves used in the analysis were inferred on the basis of knowledge of the rules and procedures used in typical plant design processes. A more rigorous analysis would include performing detailed fragility analysis for each component and combining the resulting mean component fragility curves through Boolean logic.

The tornado wind speeds of 180 mph and a straight wind of 130 mph, corresponding to 3 x 10-6 and 2 x 10-4 annual hazard probabilities of occurrence, produce damage frequencies of 1 x 10 -6 and 5 x 10 --s for core damage and production events, respectively.

For missile criteria risk analyses the simplifications are as follows:

1. Evaluation of missile damage is limited to exterior balTiers for core damage and production events.

2. Estimation of impact probabilities for vulnerable areas, such as doorways and vent openings, is not included in the analysis.

3. Failures of plant stacks due to high winds are not considered.

4. Missile generation contributions resulting from the collapse and/or failure of tall structures are not included.

5. Nonlinear relationships between the numbers of missiles, wind speed, facility layout, and damage probabilities are not considered.

For core damage events, reinforced concrete barriers were assumed to be damaged if the barriers suffered backface scabbing or worse from any single missile impact. As a first approximation, it was assumed that these minimum barrier thicknesses controlled plant design.

The sum of scabbing damage risk frequencies for minimum design requirements is 1.1 x 10-6/yr. This value is slightly higher than the core damage goal of I x 10-6/yr. However, this damage criterion is very conservative, because even if backface scabbing occurs, a core damage event is not very likely to occur. The WTWG believes that the core damage risk frequency is at least 1 to 2 orders of magnitude

less than the scabbing damage risk frequency. Therefore the WTWG believes that the recommended missile criteria conform to the core damage goal by a significant margin.

At an impact velocity of 35 mph, which is the core damage design guideline, the exceedance probability for automobile impact is 3.7 x 10-7/yr; thus, the automobile results meet the core damage performance goal.

5. SUMMARY OF DESIGN GUIDELINES DEVELOPMENT

LLNL evaluated NRC provisions and acknowledged the excess conservatism embedded in their supporting analysis. Therefore design guidelines which deviate from NRC provisions are developed on the basis of site-specific hazard assessments and EPRI report results for the project. These credible design guidelines were developed on the basis of technological advancements and past experience in wind engineering. The load specifications reflect updates and enhancements to NRC regulations. The justification and supporting bases for these updates and enhancements are provided in this paper. The proposed structural response evaluation and acceptance criteria follow industry codes and standards with minor modifications. The general design guidelines recommended in this paper are believed to closely comply with DOE orders.

Acknowledgment. This work was performed under the auspices of the U.S. Department of Energy by the Lawrence Livermore National Laboratory under contract W-7405-Eng-48.

REFERENCES

Kennedy, R. P., S. A. Short, J. R. McDonald, M. W. McCann, and R. C. Murray, Design and evaluation guidelines for DOE facilities subjected to natural phenomena hazards, UCRL-159iO, Lawrence Livermore Natl. Lab., Livermore, Calif., Oct. 1989.

Office of Nuclear Reactor Regulation, Standard Review Plan for the review of safety analysis reports for nuclear power plants, NUREG-0800 (formerly NUREG-75/087), rev. 2, U.S. Nucl. Reg. Comm., Washington, D.C., Jul,x 1981.

Office of Nuclear Reactor Regulation, Safety evaluation by the Office of Nuclear Reactor Regulation of recommended modification to the R.G. 1.76 tornado design basis for the ALWR, NRR position paper, U.S. Nucl. Reg. Comm., Washington, D.C., March 1988.

Twisdale, L. A., and W. L. Dunn, 7rna& Missile Simulation and Design Methodology, vol. I, Simulatin Methodology, Design Applications, and TORMIS Computer Code, EPRI NP-2005, Electric Power Research Institute, Palo Alto, Calif., Aug. 198I.

U.S. Nuclear Regulatory Commission, Design basis tornado nuclear power plants, NRC Reg. Guide i.76, Washington, D.C., April

Wind/TornadoD esignC riteria Development o Achieve Required Probabilistic Performance Goals1. INTRODUCTION2. DESIGN REQUIREMENTS3. DESIGN GUIDELINES DEVELOPMENT STRATEGY3.1. Developing Wind Load Specifications Based on Site-Specific Hazard Frequency Curves3.2. Developing Missile Load Specifications Based on DOE Guidelines and EPRI Tornado Missile Simulation Report3.3. Assuring Conformance of Design Guidelines by Performing Preliminary Probabilistic Risk Assessment

4. DESIGN GUIDELINES TECHNICAL SUPPORTING BASES4.1. Wind Load Specification4.2. Missile Load Specification4.3. Structural Response Evaluation and Acceptance Levels4.4. Risk Assessmenfto r Wind Speeds and Missile Criteria

REFERENCES

tornado

Documents