Thirty years after TMI: Five lessons learned

When I ask my students at the University of Pittsburgh if they know what “TMI” stands for, they respond with, “Too much information.” They can be excused; they were born long after the 1979 accident at Three Mile Island.

For older generations, of course, “TMI” will always mean “Three Mile Island.” Thus, I tell my students the story of what happened, what didn’t happen, and what we learned. In particular, I tell them that a cooling malfunction–coupled with multiple failures that interacted in an unpredicted manner and the bewilderment of control-room operators–caused part of the core to melt in the number two reactor at Three Mile Island. (The accident occurred in the TMI-2 reactor; TMI-1 is still operating today.) The reactor core was destroyed. Some radioactive gas was released a couple of days later, but not enough to cause any exposure above background levels to local residents.

There were no injuries or adverse health effects from the accident except for the immediate, short-lived mental distress produced by the accident. In fact, since the time of the initial operation of the Shippingport civilian nuclear power plant in 1957, not one single fatality has occurred as a result of the release of radiation from a commercial nuclear power plant in the United States.

There have been many independent studies to assess the possible impacts on people and the environment around TMI since the accident. None has found any illnesses such as cancers that might be linked to the incident. A federal appeals court in December 2003 dismissed the consolidated cases of 2,000 plaintiffs seeking damages against the plant’s former owners. The court said the plaintiffs failed to present evidence that they had received a radiation dose large enough to possibly cause adverse health effects. Although studies found no increased incidence of cancer as a result of the accident, they did find evidence of psychological stress. And such psychological stress will continue if we keep demonizing low-level radiation and nuclear power.

The accident at Three Mile Island has done much to shape the safety culture that permeates the nuclear power industry today. TMI was a humbling experience that changed the industry’s mind-set from one of smugness to one of increased vigilance. As noted by Angelina Howard, an executive vice president at the Nuclear Energy Institute, at Three Mile Island on the 25th anniversary, “The accident was a catalyst for positive change in the nuclear energy industry.” Each segment of the nuclear community–management, workers, designers, and regulators–was made to recognize its own share of responsibility.

We learned a lot from TMI. Before TMI, few, if any, in the nuclear industry would have believed that reactor operators would fail to recognize a stuck-open relief valve. Even fewer would have believed that operators would reduce safety injection flow in the face of symptoms of inadequate coolant. There were many lessons learned, and they can be reviewed more fully in the many good books written about the accident. Here I offer my top five:

Operator training needed to be improved. One of the most significant outcomes of the TMI accident was training improvements, which gave operators a better understanding of both the theoretical and practical aspects of plant operations and a sound basis for evaluating and responding to unfamiliar situations. Licensed reactor operator training today is conducted on full-scale replica simulators of actual plants. These simulators permit operators to practice and be tested in all kinds of accident scenarios, making them more proficient and knowledgeable reactor operators.

Sharing of industry knowledge needed to be more effective. TMI led to the establishment of the Atlanta-based Institute of Nuclear Power Operations (INPO) and its National Academy for Nuclear Training. These two industry organizations have been effective in promoting excellence in the operation of nuclear plants and accrediting their training programs. INPO has had a profound impact on the way nuclear plants are managed and operated. The proof is the steady improvement in plant performance in the 30 years since TMI. Plant capacity factors (the ratio of a power plant’s average production to its rated capability) have increased to 91.8 percent in 2007 from 58.4 percent in 1979. Meanwhile, the industry average of significant events has decreased from an average of 0.9 per year in 1989 to 0.01 per year in 2006.

Fission products don’t escape in the real world. The accident at TMI yielded insight into the “source term”–the amount of radioactive fission products released in the event of a major accident. From TMI data we learned that the release of volatile fission products was three to four orders of magnitude smaller than that provided for in the 1962 federal licensing criteria. This observation stimulated experiments at the Oak Ridge National Laboratory and elsewhere, which showed that during simulated accidents many fission products react to form aerosols that then plate out on walls or settle inside the primary containment structure. This knowledge, that strict leak-tightness of the containment wasn’t a significant factor in reducing fission product leakage to the biosphere, led to the downward revisions by the NRC toward a more realistic source term in 1995. Since that time, numerous experiments have examined the timing, magnitude, and controlling processes for fission product releases from the fuel, the primary system, and containment. Today, the magnitude of the source term available for release in an accident has been reduced significantly.

Control rooms were complex, poorly organized, and did not provide important information. Control rooms in the TMI generation of plants weren’t designed with the needs of operators in mind. Craig Faust, one of the control room operators during the event stated, “I would have liked to have thrown away the alarm panel. It wasn’t giving us any useful information.” The operators were overwhelmed and unnerved from the “alarm avalanche.” Necessary information wasn’t readily available in a convenient and understandable form. After the event, important safety system modifications were made to detect and mitigate inadequate core cooling and post-accident conditions. The next generation of reactors will have control rooms designed with human factors in mind and with computer technology that prioritizes the information operators receive.

The consequences of a nuclear accident were less than we thought. A postulated “worst accident” happened–the TMI-2 core melted. Yet, there was no “China syndrome.” And in spite of operator errors, there weren’t thousands of casualties. Similarly, the casualties from Chernobyl were largely limited in number to first responders and, except for seldom fatal thyroid cancers, far lower than what was predicted. While these lessons have been learned, we must not return to pre-TMI complacency.

Theodore Rockwell in his book Creating the New World states–and I paraphrase: “It is inevitable that as the humbling experience of TMI fades into history . . . we must assume that there will be some gradual lowering of standards until the next humbling experience. But we’ve also learned that the catastrophic vision of the impact of such an accident is wholly fictional.”

We must never forget TMI, it taught us many lessons and helped us develop a more realistic sense of what actually happens in a nuclear accident.