Learning the lessons of Fukushima

Among the lessons to be learned from the accident at Japan’s Fukushima Daichii nuclear powerplant, according to a new report from MIT, are that emergency generators should be better protected from flooding and other extreme natural events, and that increasing the spacing between reactors at the same site would help prevent an incident at one reactor from damaging others nearby.

These and other lessons are contained in a report put out this month by MIT’s Department of Nuclear Science and Engineering (NSE), based on its analysis of how events unfolded at the troubled plant in the days and weeks following Japan’s devastating earthquake and tsunami on March 13.

The specific suggestions in this report are quite different from the response by some governments — notably, Germany and Japan, which have halted or delayed expansion of nuclear power in the wake of the meltdowns and radiation releases at the Japanese reactors. In fact, the report says, health risks to the public, and even to workers at the plant, have been negligible, despite the significant releases of radiation over the last few months. There has been no loss of life associated with the accident, nor is there expected to be, the report says.

The new report, an update of a preliminary report issued in May, is available for download on the NSE website.

“A lot of information that was not available when we started has become available,” says Jacopo Buongiorno, the Carl Richard Soderberg Associate Professor of Power Engineering and lead author of the new report, which was co-authored by eight other members of the NSE faculty. However, he adds, there are some important areas where information has still not come out.

During the first days of the accident, he says, there were three critical delays that have not yet been well explained — although the report says there is no evidence at this point of any major human errors contributing to the unfolding problems. The delays involved operating some safety-critical valves, injecting water into the reactor cores and venting the containment buildings. “It’s not clear what the cause of the delays was,” Buongiorno says, but it is unlikely that these were caused by administrative delays in Japan’s control-and-command chain, as had been initially suggested.

Rather, because of the lack of power and the effects of the flooding, “there was disruption and confusion around the site” during the crucial early hours, he says. “Things that normally would take minutes, such as reading an instrument or connecting a cable or a hose, took hours” because of the lack of power and the debris and destruction. “Given the situation, they reacted as well as they could,” he says.

Among the specific suggestions the report makes:

• Emergency backup generators, needed to keep the systems running when outside power is cut off as it was in this case, should be well separated into at least two locations — one situated high up, to protect against flooding, and the other down low to protect against hazards such as an airplane crash. These generators should also be housed in watertight rooms, as they already are at many U.S. plants.
  
• In future plants, spacing between reactor buildings located at the same site should be increased — for example, by having other areas such as parking lots or support buildings in between — and systems such as ventilation shafts should be kept separate, in order to avoid a domino-like spread of problems from one reactor to another. In the Fukushima accident, it seems that hydrogen vented from reactor unit 3 may have reached unit 4 through the ventilation system, causing an explosion there.
  
• Officials should be cautious about decisions to evacuate large areas around a damaged nuclear plant in cases where the population has already been devastated by a natural disaster. At Fukushima, “ironically, the biggest [health] consequences may be from the prolonged evacuation,” Buongiorno says.
  
• More attention needs to be paid to how radiation risks are communicated to the public, rather than the confusing mix of different measurements that were disseminated in this case. The most useful standard is to relate radiation releases to natural background levels, rather than using technical units unfamiliar to most people.

But the report also emphasizes that all engineered structures — bridges, powerplants, skyscrapers, dams — have their own risks, especially when subjected to extreme conditions they were never designed to withstand. The authors suggest it is important not to overreact to particular high-profile cases.

“If you have an accident in your car, you don’t stop driving a car, you learn from it,” Buongiorno says. Continuing the analogy, “in this case, the accident was like a tree that fell on the car. It wasn’t the car itself.”

But to fully absorb and learn from the lessons of this accident may take years, Buongiorno cautions. “It took 20 years to fully absorb the lessons of Three Mile Island,” he says. “Some of these questions are complex, requiring quantitative analysis to fully evaluate the data and make rational decisions about how best to respond.”

Romney Duffey, a principal scientist at Atomic Energy of Canada Ltd., says, “This report is both an excellent summary and provides thoughtful suggestions. Of particular importance, the report addresses the links to energy policy and comparative-risk aspects, in addition to the purely technical and licensing considerations.” He adds that the suggestions regarding better public communication about risks from radiation releases are especially useful: “The concept of using easier-to-understand measures of risk is vital to better communicating with everyone during such times of great stress and uncertainty.”