Research in New Reactor Concepts : Tests Seeking Safer and Cheaper Nuclear Power
Three weeks before the Soviet disaster at Chernobyl reminded the world about the hazards of nuclear energy, a reactor in the heart of the Idaho desert went through two accidents most feared by the people who design and operate nuclear power plants.
But when this reactor lost the flow of its primary cooling system, and then its secondary cooling system, operators did not rush to prevent a meltdown by opening and closing valves, flooding the reactor with emergency coolant, or sending control rods into the super-heated core to end the nuclear reaction.
Instead, they stood back and watched. The reactor heated up a bit, then paused and shut itself down.
The events were planned by researchers eager to see how the small experimental reactor would behave under the most adverse conditions imaginable. Afterward, scientists declared their sodium-cooled reactor “inherently safe”--meaning that it will shut itself down in an accident without the help of a single moving part or any action by humans.
The research here is part of a broader national and international effort to find safer, more efficient and less costly ways of producing electricity from nuclear energy.
After a worldwide energy glut, skyrocketing construction costs, and the 1979 accident at Three Mile Island, critics of nuclear power had all but pronounced the industry dead. Now, with the Chernobyl disaster highlighting vividly what can happen when this complex technology goes awry, some skeptics are wondering whether another nuclear plant will ever be ordered in the United States.
The last one ordered--in 1978 by Commonwealth Edison of Chicago--is on hold and will not be built until sometime after the year 2000.
But reactor designers and utility industry spokesmen say the end is not here. Worldwide, more nuclear plants are planned now than a year ago. And in the United States, where energy use is growing, albeit at a slower rate than before, new electricity-generating plants will be needed within the next decade, industry experts say.
So in offices from La Jolla to Pittsburgh and at field laboratories such as the one in Idaho, engineers and nuclear physicists are working hard to design and test new concepts that they hope will ensure that some of those new plants will be nuclear powered. The U.S. Department of Energy will spend more than $300 million this year to aid the effort. And most of those involved say that the accident at Chernobyl, in a plant quite different from U.S. commercial reactors, will not change the situation much.
“We’re obviously going to learn a lot of things from Chernobyl,” said Jim Moore, vice president and general manager of power systems for Pittsburgh-based Westinghouse Electric Corp., the nation’s leading manufacturer of nuclear reactors. “We’re going to learn some things about radioactivity and dispersion. But we’re not going to learn much from a design standpoint because they (the Soviets) took such a different approach.”
Westinghouse officials believe that increased energy usage will prompt utilities to build new power stations in the northeastern, southeastern and southwestern regions of the United States by the mid-1990s--too soon for a nuclear plant to be ordered, licensed and built. But Moore said he thinks that a nuclear plant might be ordered in the “next couple of years” for operation in the late 1990s.
For that reason, Westinghouse has been working to improve its pressurized water design, which is used in 40 of this country’s 101 nuclear reactors, including both units at Diablo Canyon and at San Onofre’s Unit 1 in California. Already, Westinghouse has an agreement to make a huge, 1,300-megawatt version of the reactor for a group of Japanese utilities.
Redesign of Core
The new reactor, Moore said, will be simpler, with fewer of the crucial valves that have been a headache for utility companies to monitor and maintain. The reactor will be set in a deeper vessel so that its core will be more likely to remain covered if coolant is lost. The core itself will be more compartmentalized, making accidents easier to contain and simpler to clean up, traits that Moore said are needed for economic as much as for safety reasons. Overall, the design will be aimed at minimizing worker contact with radiation, even at low levels.
Because the new design will be simpler, Moore said, it will be as much as 10% cheaper to build than reactors in service today.
Other experts argue, however, that mere tinkering with established technology may not be enough to save the U.S. nuclear industry. Radical changes may be needed, they say.
Richard Lester, an associate professor of nuclear engineering at the Massachusetts Institute of Technology, believes that the next generation of nuclear reactors will be much smaller than the 1,000-megawatt versions built during the last decade, a change that would reduce the financial risk that utilities take when they order a nuclear plant. Lester also believes that the new reactors might be built in the centralized and controlled conditions of a factory. There, he argues, worker productivity and quality would be higher than they are in the field, where today’s reactors were assembled.
Stopping Nuclear Reaction
Finally, Lester suggests that the new reactors may need to have passive safety features that could quell the nuclear reaction through natural means, eliminating any reliance on valves that might stick or operators who can misread or ignore an alarm at a crucial moment. This change could end the need for the complex and obscurely theoretical assessments of the risk of a reactor meltdown. Many lay people find these judgments difficult to understand and thus hard to believe.
“This would also make the job of licensing a nuclear plant easier, because the licensing authority, rather than requiring a paper demonstration that a plant can cope with a series of postulated accidents, could actually conduct experiments on the real system,” Lester said in a telephone interview.
Several groups of researchers already have such reactors in various stages of development.
One, which Lester termed a “clever design,” is a Swedish idea that would immerse a conventional pressurized water reactor in a pool of cold water. The pool, treated with a chemical that absorbs neutrons and thus slows the nuclear reaction, would be kept out of the core not by valves but by the natural pressure difference between the cool pool and the hotter water inside the core. If the flow of the reactor’s primary coolant should stop, the surrounding pool of water would move toward the drop in pressure, flooding the core and stopping the reaction.
Design Still Untested
But this design is so far only on the drawing boards and has yet to be tested in any form. For now, U.S. experts see it more as a curiousity than a realistic approach. Many have doubts about whether the reactor--known as Process Inherent Ultimately Safe (PIUS)--would be reliable and able to be maintained.
GA Technologies Inc. of La Jolla faces a similar challenge in convincing utility firms that the newest design of its high-temperature, gas-cooled reactor can work. GA designed such a plant at Fort St. Vrain, Colo. That full-sized demonstration reactor, although widely acknowledged as one of the nation’s safest, has been plagued by nagging problems that have made it one of the least productive.
Al Goodjohn, a senior engineer with GA, said the company’s new reactors, similar to a design under development in West Germany, will be much smaller--140 megawatts compared to the 330-megawatt Colorado reactor. These “modular” reactors could be built in a factory and transported to the plant site, where perhaps as many as four would be set up and connected to one or two power-producing turbines.
Even if the flow of the helium gas coolant were stopped in such a reactor, Goodjohn said, it would be impossible for the ceramic-coated uranium fuel to reach a temperature high enough to make it melt. Some of the excess heat would be absorbed by the reactor’s graphite core, and the rest would be dissipated through the steel and concrete confinement vessel.
Hydrogen Explosion
The GA reactor design has the most to lose from the accident at Chernobyl, where the reactor also had a graphite core. But GA scientists and other experts say that because their reactor is cooled by gas, not water, and has ceramic instead of metal-coated fuel, the key ingredients needed to produce a hydrogen explosion like the one at Chernobyl are not present.
Still, Goodjohn said, the Soviet accident might persuade the company to design a traditional containment building to surround its reactor.
“We’re approaching it from the point of view that we could put a containment on it,” he said. “It might be that at this point it is psychologically better to do, even if you don’t need it.”
The only one of the new designs to be tested in the field is Argonne National Laboratory’s metal-cooled breeder reactor, which was put through its paces April 3 at the Idaho National Engineering Laboratory in Idaho Falls.
The breeder reactor is an old idea--the first electric power from a nuclear reactor was generated by a breeder at the Idaho lab in 1951--and several full-sized breeders have been built in France and the Soviet Union. Unlike water- or gas-cooled reactors, breeder reactors produce not only heat for generating electricity but also produce more nuclear fuel, in the form of plutonium, than they consume.
Extra Safety Factor
The Argonne reactor, called Experimental Breeder Reactor II, has taken the current technology a step further to create what its backers say is an inherently safe reactor.
Like the breeders already built, the Argonne reactor core is submerged in a pool of molten sodium, which acts as its coolant. Because sodium does not boil until it reaches 1,618-degrees Fahrenheit, it does not have to be kept under pressure. The water coolant in conventional reactors must be pressurized to allow it to reach temperatures of more than 600 degrees without boiling. If that pressure is lost, through a break in a steam line, for instance, the water coolant boils away and the core can overheat.
The Argonne reactor runs on metallic uranium fuel, which, unlike the fuel used in other breeder reactors, expands as its temperature rises. When the fuel expands, its atoms spread out, making it tougher for the flying neutrons to strike and split the atoms. When too few atoms are split, the nuclear reaction cannot sustain itself, and the reactor shuts itself down.
The first Argonne test showed that even when the flow of the sodium coolant into the reactor core is stopped, the sodium pool continues to swirl about the core, propelled by natural convection. But as this sodium heats up, and temperatures in the core rise, the fuel expands and the reaction dies a natural death.
The second test showed what would happen if the reactor lost its ability to transfer heat from the sodium pool to the part of the plant that generates electricity. Again, core temperatures rose briefly, and then the reactor shut down.
Negative Aspects
But the Argonne reactor is not without drawbacks. Its liquid sodium coolant ignites upon contact with oxygen or water--always a possibility as the sodium flows through generators to create the steam that turns the electricity-producing turbines. And skeptics point out that the successful tests earlier this year did not show what would happen if the core were completely uncovered by a rupture in the reactor vessel during, for example, a catastrophic earthquake.
Even if the reactor itself could be proven safe under all conditions, its byproduct--plutonium of the kind that can be used in nuclear bombs--is sure to provoke controversy.
The Argonne concept calls for reprocessing the plutonium into usable fuel at the site of the power plant in a special facility in which the highly radioactive products would never leave the containment structure. Argonne officials say this would all but eliminate the chance of plutonium being stolen for use in a bomb. Others, however, wonder if utilities, seeking simpler concepts, will be eager to enter the new field of fuel reprocessing.
At least two firms think they will be. Rockwell International is working on an Argonne-inspired design for a modular, 300-megawatt breeder reactor, and General Electric officials say their company could be ready to begin building a 135-megawatt demonstration reactor within three years. A commercial plant might follow 15 years later.
Using Time Productively
Bertram Wolfe, a General Electric vice president and manager of the firm’s nuclear technologies and fuel division, said his firm is not bothered by the minimum two-decade lead time it will take before one of its breeder reactors might be producing commercial power. He said researchers can use the time to find bugs in their systems and redesign the reactors before they reach the marketplace, avoiding some of the problems that plague today’s plants.
“Let’s do our homework, follow a promising concept and do it sensibly while we have time, and not in a panic program where we do things the wrong way,” Wolfe said.
Robert Pollard, a physicist with the watchdog Union of Concerned Scientists, said he doubts that General Electric needs to worry about being rushed. Pollard does not think that another plant will be ordered in the “foreseeable future”--a time period he is reluctant to define. But even Pollard, a long-time critic of the industry’s practices, said he is encouraged by the trend in research toward smaller reactors and the other changes aimed at making reactors shut themselves down naturally.
Said Pollard: “I like all these designs, from the standpoint of getting the laws of physics on your side. They can’t be violated as routinely as the Atomic Energy Act.”
