ASHINGTON, March 14 - Like the taxis in Havana, American nuclear power reactors are in heavy use, important to the economy and really, really old. The most modern was ordered in 1973.
Now after decades, four huge electric companies are expressing strong interest in new reactors, and they would like a new plant to reflect some of what has been learned of the operation.
Entergy, Exelon and Dominion have each applied for advance approval on sites where they might build reactors, although they have not committed to actually ordering one. The fourth, Duke Power, met with the Nuclear Regulatory Commission on Monday to describe how it was in the early stages of preparing an application for a reactor license, although it did not say what type it wanted to build, or where it would go.
On the drawing boards are all kinds of exotic designs, using graphite and helium, or plutonium and molten sodium, and making hydrogen rather than electricity. But the experts generally agree that if a reactor is ordered soon, its design changes will be evolutionary, not revolutionary.
The utilities are not ready for a giant technology leap; they want a plant that does what the existing ones do, but slightly better. So if new orders materialize in the next five years, it will be the mechanics and engineers who will get to show what they have learned. The physicists will have to wait.
"The pitfall is too much innovation," said Jeffrey S. Merrifield, one of the five members of the Nuclear Regulatory Commission, addressing 1,400 industry professionals at a meeting last Tuesday at the commission's headquarters. He compared new designs with the "concept vehicles" that car companies display at auto shows; buyers are drawn to them, but when it is time to buy, they pick a Ford F-150 or a Toyota Camry instead, he said.
In a telephone interview, another expert, Dr. Andrew C. Kadak, a professor at the Massachusetts Institute of Technology and the former president of the Yankee Atomic Electric Company, which helped run several reactors, laid out some practical considerations.
"If I were shopping today," Dr. Kadak said, "I would like a reactor that minimizes the dependency on active cooling systems for emergency core cooling."
It would also have some smaller changes to increase efficiency and ease of working, in the nature of interior decorating. For example, Dr. Kadak said, reactors would be a lot easier to maintain if designers stopped putting pumps and valves far from platforms and stairs.
"Give me some more space, so I can at least take this pump out without having to move three other parts," he said.
Fundamental innovations that may be introduced in the long term include the "pebble-bed," which uses fuel that cannot melt at the temperatures the reactor can achieve, and the use of liquid sodium for heat transfer, allowing operation at much higher temperatures and making more efficient use of uranium or plutonium.
Westinghouse is one of the companies trying to market a reactor, the AP1000, with more modest technical goals. It has an output of a little over 1,000 megawatts with what is called a passive approach to safety. It requires only half as many safety-related valves, 83 percent less safety-related pipe and one-third fewer pumps.
In the new design, water for emergency cooling has been moved to a tank inside the containment, above the reactor vessel. The changes will allow the emergency core cooling system to run even if all alternating current power fails, Westinghouse says.
The company is trying to sell four AP1000 reactors to China.
The AP1000 is competing with the EPR, for European Pressurized Water Reactor, a creation of Framatome of France and Siemens of Germany, which both became expert in the technology as those countries continued to build reactors after the United States stopped. Their joint venture is called Areva.
Their reactor has four emergency core cooling systems, instead of the usual two. That could help safety, further reducing the small chance that the system will not be available in an emergency. But there is a more practical reason. One cooling system can be shut down for maintenance while the reactor is running without reducing the safety margin to an unacceptable degree.
The EPR is being built now at Olkiluoto, Finland. It has a containment building designed to withstand the impact of a commercial jet, and a set of features intended to cope with a molten core in case of meltdown, preventing a "China syndrome" of a core burning through the floor and into the earth beneath. The reactor has a "core-spreading area" where the molten material would spread out and be cooled by water running above and below the area.
Approval by the Nuclear Regulatory Commission is many months away, but Areva hopes American companies will buy it because of the track record in Finland. And, executives say, they have taken full advantage of everything learned in the last few decades.
The third entry is General Electric's Economic Simplified Boiling Water Reactor, derived from its boiling water reactor design.
It is tweaked for better natural circulation in case of an accident, so there will be less reliance on pumps. A typical weak spot in existing reactors is the emergency diesel generators, but this model does not need them for emergency operations.
The reactor has a higher water inventory, as a safety measure, and eliminates large pipes below the level of the core, to reduce the chance of a leak. If the reactor shuts down automatically, then the decay heat, or heat given off by radioactive material after the reaction stopped, can be removed automatically for 72 hours, with no operator action, according to General Electric. Decay heat is what melted the core of the Three Mile Island reactor.
GE portrays its new design as an improvement on its previous evolutionary version, the advanced boiling water reactor. Peter G. Wells, who is in charge of marketing the new model, said that two of the previous versions were built at Kashiwazaki in Japan in the late 1990's. "They have 15 reactor years of proven operation," he said. Two more are under construction at Lungmen in Taiwan.
It is not certain, of course, whether anyone in the United States will order a new reactor in the next few years, although high prices for natural gas and uncertainty about what rules will apply to coal plants are creating interest.
Most nuclear advocates are expecting federal help, perhaps in the form of a production tax credit, like the one given for windmills, for the first four or five reactors, on the theory that once the first few plants are built, costs will fall and other reactors will follow, unsubsidized, with a benefit to clean air and the national economy.
Cost and construction time are only projections. David Lochbaum, who once worked as a start-up specialist for General Electric reactors and is now at the Union of Concerned Scientists, a group that frequently petitions to have plants shut for safety reasons, said that the best reactor is one that has not been built yet; its problems are still undiscovered.