A Chernobyl in the US? Not too likely, say scientists. Reactor containment buildings and US design reduce the risk

Nuclear-power safeguards adopted in the United States have made the chance of a Chernobyl-type accident happening here small, although there is room for improvement and no guarantee that a major accident won't happen. There are 100 commercial reactors in the US that produce electricity. The government operates another eight reactors to make plutonium for nuclear warheads and for research purposes.

The Chernobyl reactor produces electricity but its design is substantially different most from US nuclear power plants. In American designs the atomic chain reaction automatically stops when the cooling water is removed. In the Soviet reactor this is not the case. Instead, loss of water increases the intensity of the chain reaction, adding substantially to the danger of overheating.

Another difference is the Soviet reactor's use of graphite. Only one commercial and several government reactors use this material in the US. If graphite gets hot enough, it will burn like coal and is extremely difficult to put out. Also, when graphite is exposed to steam, it can produce a highly flammable gas.

Graphite's flamability would not have been such a liability if the Chernobyl reactor had been covered by a confinement dome, as are all commercial reactors in the US. If the Soviet reactor had been similarly equipped, it would likely have been able to withstand the explosion and smother the fire by cutting off its air supply. It would also have bottled up most of the radioactivity released by the accident, US nuclear experts argue.

(The US government's plutonium production reactors, run by the Department of Energy, do not have containment structures. But they are operated in a much different manner from power reactors and their safety features are appropriate for the manner in which they are operated, Department of Energy officials say. Nevertheless, safety inspections of at least one of these reactors have reportedly been initiated as a result of the accident.)

According to Robert Bernero of the Nuclear Regulatory Commission (NRC), the decision 30 years ago to require containment structures represents a major, philosophical difference between the US and the Soviet programs. ``We decided that the risk of an accident that would destroy part of a reactor's core and so release substantial quantities of radioactive material posed a serious public danger. So we decided to enclose them in airtight `containment' structures,'' he explains.

These structures were planned using something called the ``design basis accident.'' The idea is similar to the 50- or 100-year storms and floods used for designing bridges and dams. One of the most severe accidents experts could dream up was chosen, its consequences conservatively estimated, and the containment designed to withstand them with a prescribed safety margin.

This approach proved adequate for the 1979 accident at Three Mile Island (TMI). Although the core was destroyed and large amounts of radioactivity were released from the reactor, only minute amounts were measured in the environment. As a result, no one was seriously injured.

However, this design approach came in for severe criticism. A reactor's safety features might not be as effective against other kinds of accidents, its critics argued. In recent years, nuclear safety analysts have been trying to physically model a range of severe accidents. This is known as ``source term'' research and, according to NRC's Bernero, it is showing that the safety margins in containment structures ``are very large, but irregular.''

Until the 1979 accident, the Soviets had not thought containment buildings necessary. Their basic philosophy, explains Don Winston of the Atomic Industrial Forum, was that if you design the reactor right and build all the parts with sufficient care, then major accidents should not occur. It was a view shared by many in the US nuclear industry who felt reactors were so well designed that a serious accident was virtually impossible. But TMI punctured this conceit.

While events have shown that the US course was the prudent one, knowledgeable observers generally agree that, sooner or later, a US reactor will have another serious accident, one in which a significant portion of the reactor core melts.

``This is a dangerous technology,'' comments Bruce Babbitt, governor of Arizona and a member of the presidential commission that reviewed the TMI accident. ``It is more likely than not that there will be a core meltdown in the US within the next 15 to 100 years. That is the reality we have to deal with,'' he comments.

Whether such an incident, should it happen, bears a greater resemblance to Three Mile Island or Chernobyl will depend primarily on the adequacy of the reactor's containment systems.

Although more work remains to be done, recent source-term research strongly suggests that current containment systems have a greater margin of safety than had been previously estimated. Last year, the American Physical Society issued the first independent overview of this work. Chairman of the study Richard Wilson, of Harvard University, characterized it as ``slow-developing good news for nuclear safety.''

The antinuclear group, the Union of Concerned Scientists, is one of those which argue that a Chernobyl-like accident is all too possible in the US. Explains UCS lawyer, Ellen Weiss: ``From time to time these events happen. They serve to remind us of the inherent dangers of nuclear technology.''

The organization has done its own report on source-term research. It criticizes the completeness of the work, pointing out that many potential accident sequences have not been examined. UCS critics also argue that a number of outstanding questions regarding the adequacy of today's containment systems remain. They contend that chemical explosions possible during an accident might breach them. They also report cases where reactor operators have left containment doors open for long periods of time.

The likelihood that another severe accident will occur -- as distinct from its probable consequences -- depends heavily on how well reactors are operated.

One generally accepted message from Three Mile Island was that commercial nuclear reactors are substantially more demanding to operate than other types of electrical generating equipment. And, while some of the 50 utilities with reactors were capable of building and operating them properly, many utilities were in over their heads. Since then there have been significant improvements in the operating records of US reactors, according to a variety of observers. But there is serious disagreement about whether the improvements have been enough.

The Institute of Nuclear Power Operations is a self-policing industry group formed after TMI to evaluate and improve reactor operations. ``We have seen an extreme reduction in the number of `significant events' that have been occurring. Between 1980 and 1985 we have seen a 100 percent reduction,'' reports William F. Conway, INPO's group vice-president for industry and government relations.

The group has gathered a number of other statistics that corroborate the argument that there have been significant improvements. For instance: the number of unplanned shutdowns has been reduced from 6 to 4.3 per year per reactor; average worker radiation exposures have been cut by a third to a half; and accidents serious enough to cause loss of time on the job have been reduced by 200 percent.

Even James MacKenzie of UCS acknowledges that ``there is no question that overall safety has improved since Three Mile Island.'' Unlike industry sources, however, he does not consider these improvements adequate.

Mr. MacKenzie and a number of others concerned with nuclear safety are highly critical of the job the NRC has been doing. ``NRC is in worse shape than it was eight years ago,'' says governor Babbitt.

In a recently released report, the Nuclear Regulatory Commission said safety suffered at commercial nuclear reactors in the US last year. It cited inadequately trained operators, broken emergency equipment, and poorly maintained equipment. ``This past year has been the worst year for nuclear safety in the domestic nuclear industry since [TMI],'' said Rep. Edward J. Markey (D) of Massachusetts, chairman of the House energy conservation and power subcommitte which has been looking at the Chernobyl accident.

One suggestion made by the TMI commission was to replace NRC's five-man commission with a single administrator. The idea never caught on in Congress, but appears to have been revived by the Soviet accident: At a hearing last week on Capitol Hill, both NRC commissioners and industry spokesman testified in favor of such a change.

In addition, bills have been introduced in both the House and Senate to establish an independent board to review incidents at nuclear reactors. It would be analogous to the National Traffic Safety Board that investigates airplane crashes. The measure stems from concerns about the adequacy of NRC's previous procedures.

Last year, Walter Kato headed a study of NRC's accident investigation process. ``We found a number of areas of possible improvement,'' he reports. Specifically, they discovered a lack of coordination in NRC investigations and a tendency to look for violations of regulations rather than determine the cause of problems. Also, they uncovered cases where significant recommendations were never acted upon. The group recommended that NRC form a special investigation unit, which NRC did.

Currently, a much different approach to increasing the manageability of nuclear power is also being pursued: the development of ``inherently safe'' reactors. These are designed so that it is physically impossible for the core to melt.

The foremost proponent of such reactors is the Gas-Cooled Reactor Association. According to GCRA's Archie Kelley, they have adopted a reference design with four 135 megawatt reactor modules driving two steam turbines. These are designed so that the core will not overheat even if the reactor loses all its helium coolant.