In March, 1979, leaking coolant at Three Mile Island Reactor Unit 2 in Pennsylvania brought the nuclear power plant within 30 minutes of a catastrophic meltdown. The accident was contained, and only small amounts of radioactive gas escaped from the damaged reactor.

The near disaster marked the beginning of the end of America's romance with atomic fission. Yet it didn't curtail the growth of nuclear power. Indeed, ever since Three Mile Island, efficiency improvements have helped nuclear-power generation to grow steadily (chart), even as the number of functioning commercial reactors has fallen to 103 from a peak of 109.

And there's more to come. Over the next decade, nuclear output will grow an additional 10%, says Tom Christopher, CEO of Framatome ANP Inc., the U.S. unit of Paris-based Framatome ANP, the world's largest provider of nuclear engineering services. The result will be an additional 10,000 megawatts' worth of electrical capacity--the equivalent of 8 to 10 big nuclear facilities--without requiring the construction of a single new plant.

The surge in capacity growth, Christopher says, is a result of the ongoing relicensing of the nation's commercial nuclear fleet. Today's power plants were commissioned to split atoms for not more than four decades. Starting in the late 1990s, the Nuclear Regulatory Commission (NRC) began to extend that term to 60 years on a plant-by-plant basis. The extensions have opened the door to major capital investment, much of which is funneled through Framatome. Plants pay the engineering company to upgrade their key systems, replacing clunky, 1970s-era generators, mechanical switches, and manual gauges with high-efficiency motors and digital controls. The upshot is not just increased output but improved safety, says Christopher, a 29-year veteran of the nuclear-power business.

Industries Editor Adam Aston met with Christopher to learn more about the surprising growth of the nation's nuclear capacity. Excerpts from their conversation follow.Q: Why have the operating extensions made such a difference to the industry?A: Under NRC guidelines, the operators can submit an application for a renewed license within three years of a facility's 30th year. The renewal adds 20 years to the plant's original 40-year license. The life extensions open the door to capital improvements and make it possible for operators to take advantage of the lessons learned over the past 30 years, to retool and upgrade for another 30.Q: What kinds of changes are taking place?A: Every year, our ability to upgrade a plant improves. Productivity gains have been so high over the past 20 years that our costs to upgrade a plant have fallen by half. For instance, when a nuke refuels, every 15 to 18 months, it's required to do an intensive inspection. In the past, that was done by people--even in high-risk radioactive areas. Now in practically every instance, we have a robot do the work. These machines can even do repairs--they can weld and grind. So now, labor counts for only about 30% of the value we provide in an upgrade.Q: What is the scale of these upgrades?A: It depends on what the operator is willing to spend. And that, in turn, depends on the average price of power over the plant's remaining years. If you assume a conservative price--say $2.50 to $3.50 per megawatt hour--a typical facility could justify $100 million to $200 million in spending per reactor and still recover that over 20 years.Q: How do you decide what to replace?A: The majority of U.S. plants were designed in the late '60s and '70s. In many ways, they're crude by today's standards. But they were designed very conservatively, with lots of redundancy, so there are parts that don't need to be changed. Also, it varies with the unit. If the plant is on a lake and cannot increase its discharge of cooling water, then upgrading its generating capacity isn't an option. If a plant is able to boost its output, then we can replace the steam turbines and generators.A lot of little things can also increase efficiency--and power output. There are thousands of detectors in a nuclear plant measuring things like temperature and pressure. Each is connected to an electromechanical control panel. You can replace those analog detectors and gauges with microprocessors that will do more. And you can integrate the controls into a simpler system that requires fewer engineers to monitor.We can also reduce the house load power--the electricity the plant needs to operate. The cuts can be significant--say, 40 to 60 Mw. It's not unusual for a plant to have 3,000 motor-operated valves. We can replace these valves and pumps with more efficient variable-speed motors, cutting the house load by 10%. And all that power can be sold to market.Q: What's the net effect of these upgrades?A: You will hear industry people say we've begun a period of pseudo-construction of new nuclear plants in the U.S. On average, we'll see a 10% capacity increase from the nuclear plants here, so you're talking 10,000 Mw in the next 10 years.Q: What sorts of efficiency gains have we already seen?A: Think of it in terms of capacity factor, which is the industry's actual production as a percentage of its potential maximum. The average for the U.S.'s 103 nuclear plants is 91%, the highest such rating in the world. It means that a typical plant is down only 9% of the year, or 33 to 35 days. That's remarkable, especially since in the early 1970s that measure was 60% or so--around eighth place compared with other national nuclear fleets. The improvement began before the current round of relicensing. It's due partly to the efforts of the industry associations to share operating practices.Q: Will the U.S. build any new nuclear plants?A: Given the volatility of power prices, nuclear operators look at the near term--say, three to five years. In that time, is anybody going to need a big base-load nuclear plant [i.e., a large-capacity facility that is run continuously]? Not likely. In 5 to 10 years, there may be a window. If so, the decision will probably be driven by other issues, such as environmental constraints. You might see the value of nuclear facilities rise if the world moves toward some sort of a carbon tax. Since it emits no greenhouse gases, nuclear power could be used to offset dirtier sources. It's impossible to predict what sort of energy technology will be available then. Perhaps we'll have a hydrogen economy where nuclear power will be used to split water into hydrogen gas.Q: What are your thoughts on radioactive waste?A: It's important to put the problem in context. I've seen data that say if you take all of the spent fuel rods generated in nuclear plants in the U.S. and stack them up, you'll have a pile that's 10 yards high and fits inside a football field. That's it. Now, the issue is how do we deal with it. To us, Yucca Mountain [Nev.] is the ideal solution. And frankly, the tax that's currently in the electricity rates--two-tenths of a cent per kilowatt hour--would be more than enough to build and operate Yucca Mountain. All of the nuclear plants today have established a decommissioning reserve in cash.Q: In 1998, Germany voted to phase out its existing nuclear plants. Does this mean few nukes will be built abroad?A: Some countries are backing away from nuclear energy. But the news is more positive than negative. Finland just approved a public referendum to build a new nuclear plant. And in the former Soviet Union, they are determined to go back and complete a number of their plants that were never finished. Framatome is completing work on two plants in China. South Korea, of course, also continues to build nuclear plants. And Japan has a robust construction program--maybe six or eight more plants are planned over the next decade.