In 2007, the first application to build a new reactor in the United States in more than three decades was filed with the Nuclear Regulatory Commission (NRC). By the end of that year, four more applications had landed at the agency. In 2008, 12 additional applications arrived, with one more filed in 2009. Nuclear backers proclaimed a “renaissance” underway.
The NRC, which over the years had lost personnel because of a shortage of work, geared up, hiring 1,000 new staffers to handle the licensing requests. Things got so crowded at the Office of New Reactors that in May the agency broke ground for a third office building in suburban Washington.
A new generation of nuclear power is indeed taking shape, driven, in large part, by a growing sense among environmentalists and policymakers that any strategy to wean the U.S. off planet-warming fossil fuels must include construction of more nuclear power plants. But how safe will this new generation of nuclear power plants be in comparison to the existing fleet of 104 plants that currently generate 20 percent of the nation’s electricity?
Perhaps the most critical difference is that the new designs are simpler and rely less on human or mechanical intervention in the case of accidents. Settling on a standard design was one recommendation made after the 1979 accident at Three Mile Island. Some designs, for example, use gravity to provide emergency cooling water rather than pumps, which can fail. Some reactors now have redundant safety features, like extra pumps. In addition, the NRC has increased regulatory scrutiny of the new designs, ordering, for example, additional safety features or engineering changes to improve delivery of emergency cooling water.
Russ Bell, director of new plant licensing at the industry’s Nuclear Energy Institute in Washington, maintains that the new plants will be extraordinarily safe. Government risk assessments for the new reactor designs say that an accident that could damage the reactors’ cores would likely occur once every 10 million years — an order or two of magnitude lower than the U.S’s existing nuclear power plants. And even were a core-damaging accident to occur, Bell says that does not mean radiation would escape, since the reactors have containment buildings and systems designed to prevent releases of radioactivity.
Since the terrorist attacks of Sept. 11, 2001, all containment structures on new nuclear plant designs in the U.S. have been re-engineered to withstand the direct impact of a jetliner. This does not mean, however, that new containment designs are foolproof. The containment structure on a popular Westinghouse design, which seven utilities are now considering building, has been upgraded, but the NRC determined it probably won’t withstand a severe earthquake.
It’s also worth noting that the NRC does not require the new plants to be any safer than existing ones. Rather, it only requires the plants to “provide the same degree of protection” as the current generation of reactors.
The new reactors remain a work in progress. Even without knowing exactly what the finished reactors will look like — or cost — some utilities have already made their choices, spurred on by promises of federal subsidies and political pressure to cut carbon emissions. In a speech to industry leaders in May, Nuclear Energy Institute CEO Marvin Fertel said that the construction of nuclear reactors to provide additional power and to replace older plants — U.S. reactors are limited to 60 years of operation — means that 187 new nuclear power plants must be built by 2050.
Many outside the industry believe that figure is unrealistically high.
Elsewhere around the globe, nuclear power expansion is underway. Today, 436 reactors are operating in 31 countries, generating about 15 percent of the world’s electricity. Fifty reactors are under construction, primarily in China, South Korea, and Russia, with the fastest growth in Asia. India, France, and Finland also are building new plants.
Most experts don’t expect a new reactor operating in the U.S. before late 2016 or early 2017.
Although the new U.S. reactors will have some “design enhancements” — digital controls versus analog dials, for example — “at bottom they are based on familiar and proven technology,” says Bell. The two underlying technologies — pressurized water reactors and boiling water — have been around since the start of the nuclear power era. (In a pressurized water reactor, superheated water is pumped under high pressure to the reactor core; the heated water then transfers its thermal energy to a secondary steam system that turns a turbine to generate electricity. In a boiling water reactor, the water is injected directly into the core, creating a water-steam mixture that turns the turbine. Most reactors in the U.S. are pressurized water reactors.)
Those technical similarities are a good thing, according to nuclear safety watchdogs. “The further away you are from systems in common use and from actual construction experience, the bigger the uncertainties are going to be,” says Edwin Lyman, senior scientist with the Union of Concerned Scientists, which has been evaluating reactor safety for four decades.
Under ideal circumstances, there would be just one or two designs under consideration, not five. Settling on a “standardized” design was among the recommendations made by nuclear advocates and critics alike in the aftermath of 1979’s accident at Three Mile Island in Pennsylvania. Having just one or two “off-the-shelf” designs would simplify licensing, construction and operation, and hold down costs. This country’s existing reactors are custom-made; no two are alike, which means they are extremely complex to build, run, and regulate.
But companies hoping to dominate the U.S. market have filed applications to build a variety of designs, and the NRC has committed to reviewing the massive documentation for each.
For many at the NRC, this is new work: Half the agency’s workforce has been on the job for less than five years. And the information provided by the manufacturers is sometimes lacking. Last year, NRC chairman Gregory B. Jaczko complained that “we have incomplete designs and less than high-quality applications submitted for review.”
It will be at least 2012 before the first new design wins final approval from the NRC. The four other designs are lined up behind that on the NRC’s calendar, pushing licensing into the middle of the decade. Indeed, the approval process already is behind schedule because of safety issues with some reactor designs, such as the integrity of the containment dome around the AP 1000 design from Westinghouse.
Most experts don’t expect a new reactor to be operating in this country before late 2016 or early 2017.
One reason utilities are committing to new construction is to snag financial inducements from Washington.
The pace of design reviews and licensing contrasts sharply with the political push to build new nuclear plants, which are regarded by many on Capitol Hill and in the White House as key to combating climate change. That has created the curious situation in which utilities have announced plans to build reactors from specific vendors before they know everything about what they’re buying. Part of the reason utilities are committing to new construction now is to snag attractive financial inducements from Washington that are being offered on a first-come, first-served basis.
In recent months, the Obama administration and nuclear backers in Congress have beefed up incentives first offered in the 2005 Energy Policy Act. In February, the White House announced $18.5 billion in tax credits, as well as loan guarantees for new reactors. The Kerry-Lieberman climate bill would raise the guarantees to $54 billion, and some in Congress favor no limits. (The loan guarantees are regarded as critical to help utilities cut their borrowing costs for the first new reactors, each of which is expected to cost $10 billion to $12 billion.)
What are the designs under review and what new features do they include?
One important distinction is whether the reactors rely on “active” emergency cooling systems (which depend on mechanical equipment and uninterrupted electrical power supplies) or “passive” systems (which rely on gravity or other natural features). Current reactors utilize active systems. Adoption of passive systems was ranked high on the list of recommended safety changes in the aftermath of Three Mile Island, although passive systems could pose risks, such as being unable to supply enough water where it’s needed, when it’s needed.
When risks from external events are factored in, the new reactors are no safer than older ones.
There is very little actual experience in either the construction or operation of the new reactor designs to guide utilities in making their choices. “There doesn’t seem to be much difference in the price ranges of the various designs,” says Lyman. “But a lot of the cost will have to do with the learning curve of building new reactors again.” (There have been no new reactor construction starts since 1977.)
Will this new generation of reactors be safer than the current nuclear plants? Ask that of the industry’s Russ Bell and he chooses his words carefully, because to imply that the new reactors are “safer” than the old ones infers that the existing plants are less safe. “We think all our plants are safe,” he says.
The industry has performed complex mathematical analyses called probabilistic risk assessments to measure the likelihood of a serious accident, says Bell. “When you run the numbers on the newer designs, as you’d expect, the chance of a damaged core or release of radiation accident is much, much lower than the current fleet,” he says. “But the numbers are very low for all the plants.”
The analysis is limited, however. Computed risks for new reactors are lower than for current designs “when only internal events are considered,” according to a 2009 report that the Nuclear Energy Institute wrote for the NRC. (That includes fires or pipe breaks, for example.) But when risks of damage caused by external events — earthquakes, for example — are factored in, the new reactors are no safer than older reactors. In addition, because utilities have no operating experience with the new reactors, the probable risk assessments are purely theoretical and not as reliable as years of actual operating data from existing plants.
While the NRC continues its evaluation of the five reactors, Lyman argues that none is as safe as it could be. The new designs are engineered only to withstand a predictable sequence of events, something engineers theorize may happen. In nuclear parlance that is called a “design basis accident.” The new reactors, like their older counterparts, are not designed to survive an unexpected sequence of events. That is the critical flaw, says Lyman: “Three Mile Island was a beyond-design-basis accident.”
Editor’s note, June 17, 2010: An earlier version of this article incorrectly stated the basis of British and Finnish concerns about Areva’s Evolutionary Power Reactor. The British and Finns raised concerns about the reactor’s control and instrumentation systems, not about the reactor’s containment integrity.