Tornado Hazard Assessment

DOE-STD-1020-2002

The following argument is presented to justify the uniform approach to wind design.

ASCE 7 addresses the physical characteristics of wind, including variation of wind speed with

height and terrain roughness, effects of turbulence, and the variation of wind pressure over the

surface of a building. Wind effects addressed in ASCE 7 can be detected and measured on wind

tunnel models and on full-sized structures. Furthermore, evidence of the physical effects of wind

found in wind tunnel and full-size measurements are also found in windstorm damage. The

appearance of damage from straight, hurricane and tornado winds is very similar. The similarity

suggests that wind pressure distribution on SSCs is generally independent of the type of storm.

One cannot look at a collapsed windward wall, or an uplifted roof, or damage at an eave or roof

corner or wall corner and determine the type of windstorm that caused the damage. Table D-1

lists specific examples where the appearance of damage from the three types of windstorms is

identical. Many other examples could be given. The conclusion reached is that the proposed

uniform approach is reasonable for estimating wind loads produced by straight winds, hurricanes

and tornadoes.

D.2

Tornado Hazard Assessment

The traditional approach for establishing tornado criteria is to select extremely low

exceedance probabilities. The precedence was established in specifying tornado criteria for the

design of commercial nuclear power plants. An annual exceedance probability of 1x10-7 was

adopted circa 1960 when very little was known about tornado effects from an engineering

perspective. Much has been learned since 1960, which suggests that larger exceedance

probabilities could be adopted. Some increase over the 1x10-7 is justified, especially for

facilities that pose substantially smaller risks than commercial nuclear plants. However, two

factors make it possible and desirable to use relatively low tornado hazard probabilities: 1)

straight and hurricane winds control the criteria for probabilities down to about 1x10-4 and 2)

additional construction costs to achieve low tornado probabilities are relatively small, when

compared to earthquake design costs. For these reasons, the tornado hazard probabilities are

set lower than straight winds and hurricanes. They also are set lower than earthquake and flood

hazard probabilities.

D-2