Recently an outbuilding at a nuclear power plant received a glancing blow from a tornado. Fortunately no real harm was done, but it started me wondering: are nuclear power plants built to withstand a direct hit from a tornado?
Illustration by Slug Signorino
There’s a range of possible answers to this question:
1. Yup, 100 percent guaranteed.
2. Hope so.
3. Oh, shit.
No one can ever honestly give answer number one. Nuclear-power engineers like to think they can use answer two without crossing their fingers. However, they thought the same thing at the Bureau of Underwater Oil Well Leaks.
The close encounter with a tornado you’re probably referring to involved the Fermi 2 nuclear plant in Michigan. Although the reactor shut down due to a partial loss of emergency backup power, actual physical harm was limited to a hole in the roof, siding stripped from an outbuilding, and some damage to the cooling tower, which is actually less scary than it sounds.
Tornado-related structural damage comes from three sources: the wind itself, suction (i.e., sudden drops in air pressure), and flying debris. In the early Atoms for Peace days, the Atomic Energy Commission merely required that plants be able to withstand high winds, but in the late 60s regulators began thinking harder about suction and debris.
To get a better handle on how bad tornadoes could get, the government looked at the research of Ted Fujita, creator of the F-scale of tornado intensity, which rated twisters from F0 to F5 based on the damage they caused. (Instead of these dull numbers, your columnist prefers Fujita’s original terminology, which classified levels of damage as “devastating,” “incredible,” and “inconceivable.” But one recognizes these terms detract from the requisite veneer of seriousness.)
In 1974 the first major regulations for tornado-resistant design came out, requiring that nuclear plants in most of the U.S. be capable of surviving a total wind speed of 360 miles per hour — a figure that was literally off the charts, as the F-scale topped out at 318 mph. That raised the question of how tornado-resistant pre-1974 plants were. A mid-70s study of nine early plants found the odds of serious tornado damage in any given year were less than one in 5 million, with damage likely limited to the backup power systems. The chance of a tornado-induced core meltdown was calculated at 1 in 15 million over a reactor’s 30-year life span.
To the jaded modern ear, those numbers may sound too reassuring to be right, and in fact research established that severe damage can occur at much lower speeds than Fujita initially thought. This gave rise to the Enhanced Fujita scale, or EF-scale, introduced in 2007, which greatly lowered estimated wind speeds for the most destructive tornadoes (EF3 and higher).
The current design standard requires that nuclear plants be able to withstand “the most severe tornado that could reasonably be predicted to occur at the site,” based on a study of more than 50 years of tornado data. Today nuclear plants in the midwest and Great Plains must be designed for total wind speeds of 230 mph, which isn’t a relaxation of the earlier standard but rather reflects a better understanding of how much damage can occur at that speed.
Details of the current standard are frighteningly but somehow reassuringly practical. A nuclear plant must be able to safely survive the impact of a one-inch steel ball hurtling through the air at 17 mph, a 15-foot length of six-inch-diameter steel pipe flung at 92 mph, and a 4,000-pound car flying at the same speed.
What kind of tornado damage have nukes suffered to date? Nothing that came close to releasing radiation, although buildings and equipment have certainly gotten roughed up some. The first incident occurred at the Grand Gulf Nuclear Generating Station in Mississippi, which encountered an F3 tornado on April 17, 1978, while the plant was still under construction. Damage was limited to the electrical switchyard and a cooling tower, which lost a big chunk of concrete from the top.
In 1998 the Davis-Besse Nuclear Power Plant in Ohio was hit by an F2 tornado, which damaged the switchyard and communications and forced the plant into automatic shutdown after external power was lost. (Arguably more damage resulted from lightning strikes than winds.) Due to the lack of power, a spent-fuel storage pond got warmer than the operators would have liked, but no radiation was released.
On August 24, 1992, Hurricane Andrew, then a Category 4 storm (equivalent to an EF2 or EF3 tornado), caused extensive but ultimately minor damage to the Turkey Point Nuclear Generating Station in Florida. The reactor shut down following loss of outside power and phone systems plus damage to the fire protection systems, emergency generator, and several outbuildings.
The Nuclear Regulatory Commission seems anxious to demonstrate that it’s not taking a casual attitude toward these things. In 2009 it rejected the Westinghouse AP-1000 reactor design — regulators feared the shield building, with walls of steel and concrete three feet thick, might not be strong enough.
Send questions to Cecil via firstname.lastname@example.org.