SAVANNAH RIVER SITE, S.C. — Scattered among the pine forests of this 310-square mile federal reservation are five mothballed nuclear reactors where tens of thousands of workers were once engaged in a grim race to create as much plutonium as they could.
By the time production ended here in 1988, the site was a horrendous mess. Today, about 36 million gallons of radioactive liquid wastes sit in underground tanks, and most federal spending here is devoted to nuclear-related cleanup, not production.
So when U.S. and Russian dignitaries in 2005 helped turn the first shovels of dirt here for a new plant to produce reactor fuel from weapons plutonium, it seemed like an entirely new industry, with a vital international purpose, was coming to Savannah River.
Eight years later, though, that optimism has given way to exasperation.
The fuel plant now under construction was the product of a U.S. agreement with Russia for each country to dispose of 34 tons of its own plutonium, an explosive material withdrawn from retired nuclear weapons. The plant was supposed to transform the plutonium into fuel rods to produce electrical power, ensuring it could not be put back in weapons or stolen by terrorists.
So when a reporter asked one of the shovelers at the ceremony, Sen. Lindsey Graham, R-S.C., if Congress would pay for the expensive facility, he said all other threats "pale in comparison of what could be done if plutonium fell into the wrong hands.”
Since then, the cost of the U.S. plant has ballooned while its goals have contracted.
The original agreement with Russia has been changed so significantly that it may help Russia produce more plutonium, not less. And today — after a federal expenditure of $3.7 billion — the incomplete facility is on the brink of being cancelled. “Cost growth and fiscal pressure may make the project unaffordable,” states the Energy Department’s latest budget proposal to Congress.
The project’s steep decline from a presidentially-blessed priority to a multi-billion dollar concrete edifice that may never produce an ounce of fuel is due in part to a stubborn gap between the Energy Department’s predictions about the plant and the messy reality of its design and construction.
A close look at its history reveals that the project had a weak construction contract; that Washington officials initially paid little attention to the project, despite its scope; that mismanagement and shoddy contracting practices cost the government more than $1.38 billion in avoidable expenses; that its piping may be flawed; and that prospects for selling its plutonium-laced fuel appear dim.
In 2002, the Energy department said the mixed oxide or MOX plant would cost $1 billion to design and construct and start operating in 2007. But the lead contractor — Shaw Areva Mox Services LLC in Baton Rouge — told the Department of Energy last August the true cost was likely to be $7.37 billion, a fact that DOE kept hidden from the public for months.
Since then, the DOE has told auditors the construction will cost at least $7.7 billion and it won't begin operations before 2019, about 12 years late. The cost of keeping it running over 15 to 20 years is likely to be another $8 billion to $12 billion, bringing the total price to nearly $20 billion.
Matthew Bunn, a nuclear expert at Harvard, said that according to the government's own figures, using MOX to eliminate excess U.S. plutonium could thus cost around $243,000 a pound. Disposing of extremely hazardous industrial waste, in comparison, costs thirteen cents a pound. These figures assume that the MOX plant isn’t stopped dead in its tracks by cost over-runs and the current federal budget squeeze.
The steep pricetag is partly due to the grave risks of handling plutonium, a potent carcinogen when it is inhaled and prone to explode if concentrated or compacted. The plant is designed to take this dangerous material, once it is ground up at other Energy Department facilities, and mix it with uranium oxide, the fuel that nuclear power plants typically burn.
This mixed oxide, or MOX powder, would then be compressed into pellets, baked and hardened in furnaces, and machined to a precise size to fit inside nuclear fuel rods. The rods would be loaded into nuclear reactors, where the plutonium would be partially burned to heat water that drives steam turbines.
Unexpected challenges arise
Optimism that this would be straightforward was widespread at the outset — in contrast to today, when the contractor has warned employees to expect layoffs and slowdowns.
Energy officials told Congress in 2002 that the design of the plant was 60 percent complete and its technology had been proven by decades of operations by two MOX facilities in France. Areva, which owns 30 percent of contractor Shaw Areva MOX Services LLC, is a French government-owned, nuclear behemoth with more than 46,000 employees that built more than 100 operating reactors and has a huge financial stake in pushing related technology.
But by the time of the 2005 groundbreaking — which was mostly for show since actual construction did not start until two years later — the MOX project was in deeper trouble than most people knew.
Its designers had discovered that using weapons-grade plutonium, instead of the plutonium produced by civilian reactors like those in France, posed special challenges, and so by then the project had eaten up 45 percent of its budget while the design was still only 70 percent complete.
It had reached this point without establishing key operating requirements for the completed plant, a circumstance that alarmed Energy Department Inspector General Gregory Friedman when he reviewed the program in the summer of 2005.
Friedman’s report disclosed that at the outset the NNSA had assigned only a single technical person to the project and that by 2004 only a “handful” of employees were overseeing it. No one from NNSA was stationed at the site until around 2003, even though the contractor was spending $13 million in taxpayer funds every month. Monthly reports on the work were confusing and misleading, and budgets lagged months behind the actual work, Friedman said.
A $39 million cost ceiling on one aspect of the project was somehow set 15 months after the contractor had already spent $47 million to complete the work, his report said. The Defense Contract Audit Agency, asked to examine two contract change proposals in 2003, withdrew from the assignment because it could not find a decent rationale for them in the submitted documents. And the NNSA regularly approved spending increases without demanding corrective action, Friedman wrote.
The MOX project contractors had little incentive to find efficient solutions, because NNSA had signed a “cost-plus fixed-fee” deal in which the contractor passes on to the government whatever costs it incurs — even though the agency was warned against doing so by a review team in 1999. In his 2005 report, Friedman concluded the real cost of the MOX plant was likely to be $3.5 billion, or more than triple the initial estimate.
The initial design, moreover, proved flawed: It turned out that the facility needed nearly 1,400 miles of electrical cabling, not 735 miles; it also needed engineering supports that were stronger and more costly to install than anticipated.
So in April 2007, the estimated cost jumped again, this time to $4.8 billion, according to an internal DOE memo. Meanwhile, the government had started scaling back the project’s scope.
The projected cost of a related plant meant to tear apart and grind up plutonium weapons cores more than doubled, so it was cancelled and the work parceled out to existing DOE facilities. A Government Accountability Office report partly blamed ineffective DOE oversight and poor contractor work. The Senate Appropriations committee said it supported the decision but expressed frustration that NNSA had failed to “identify alternatives earlier, before spending $700,000,000 over 13 years and determining that existing facilities could be used to meet mission needs.”
The MOX fuel facility itself, the heart of the disposal effort, was also scaled back. Instead of being equipped with three furnaces to harden the fuel pellets, the plant is to have only two. One of three presses meant to mold the fuel pellets was dropped, as were two of four emergency diesel generators.
Kelly Trice, president of Shaw Areva MOX Services, a partnership between Areva and the Shaw Group that manages the MOX construction work, has attributed the cost overruns mostly to the difficulty of finding and retaining qualified workers at a time when other nuclear plants were being started in the Southeast, and the need to get regulatory approval for “every piece of equipment, every piece of pipe, the concrete, even the dirt under the facility.” Speaking with reporters at an industry conference on Feb. 22, he also cited the rising cost of fuel and materials, including a spike in the price of concrete caused by the 2008 Beijing Olympics.
Faulty rebar undermines the construction
But some of the project’s problems resulted from cutting corners, according to an April 2009 report by DOE Inspector General Friedman.
Soon after construction started, when someone at the construction site attempted to hammer a steel bar important to the reactor’s safety into a needed shape, it broke instead of bent. That small but shocking event set off a major investigation leading to Areva’s return of 935 tons of rebar, supplied by a subcontractor whose work had not been inspected and didn’t meet federal standards for nuclear construction. Another 135 tons of improper rebar were diverted to other uses and 14 tons already embedded in concrete were allowed to remain in place after special inspections by the Nuclear Regulatory Commission (NRC). Fixing the problem cost more than $680,000.
Three years later, Shaw Areva fired construction supervisor Jeffrey Derrick of South Carolina after he complained about inconsistent or missing instructions for the installation of the plant’s 80 miles of piping. In a whistleblower complaint Derrick filed with the Energy Department, he alleged that half of the work might have to be ripped out and redone. “This will be costly and time consuming….We can not [sic] continue to ignore this issue,” Derrick told a Shaw Areva vice president on March 15, 2012, six days before Derrick says he was fired. DOE’s Office of Hearings and Appeals, in a ruling last October, rejected Shaw Areva’s bid to have the complaint dismissed. Derrick declined to discuss it, and the NNSA did not respond to a request for comment.
Separately, on April 29, Shaw AREVA reported to the nuclear commission that it had taken delivery of some defective corrosion-resistant, stainless steel piping, used in what the industry calls “high-stress environments,” leaving them susceptible to corrosion by acids found in the MOX process. The company didn’t say how much piping was affected, but told the NRC that five different sizes were involved and that it was still testing spools of pipe to determine the problem’s scope.
In 2010, the commission had faulted Shaw Areva for not identifying or correcting “conditions adverse to quality” at the site and in its equipment. Shaw Areva did not contest these claims, which it described as minor. But it also became embroiled in a dispute with the commission over whether it should have filed annual reports to the agency listing all of the 18,000 design changes made to the plant since construction began.
Then, last month, Friedman again faulted the project’s managers, this time for approving millions of dollars in “temporary living expenses” for workers who did not qualify.
Areva declined to address the construction problems in detail, but spokeswoman Kelly Cousineau said in a statement that “MOX is one of the first NRC-regulated, new nuclear construction projects in the U.S. in over 20 years; helped to revive a dormant nuclear industry by standing-up a national supply network; [and] has reached nearly 60% completion of major facilities.”
No customers for the fuel
Neither the contractor nor the Energy Department have been able to resolve another problem, however: Although the plant is supposed to make nuclear fuel for reactors, no reactor operators have agreed to buy it. That could leave the plutonium in unused fuel rods, where it could more readily be diverted to dangerous uses.
Heightened safety concerns pose one challenge. After the March 2011 tsunami crippled a MOX-burning reactor at the Fukushima Daiichi nuclear power plant in Japan, Edwin Lyman, a physicist at the Union of Concerned Scientists, a group skeptical of nuclear power, calculated that the increased plutonium released by a major accident involving a reactor using MOX could produce up to 30 percent more deaths than a reactor without it. The Energy Department has said the increased risk would be much smaller than Lyman calculates.
Unexpected technical problems have also arisen. Duke Energy was interested in using the fuel until 2005, when it was granted a license to burn French-made MOX in a trial run at its Catawba Nuclear Station in York County, S.C. The company abandoned the testing after part of the structure holding the fuel in the reactor core expanded more than expected.
Since then, Shaw Areva has been unable to find a utility willing to burn the mixed-oxide fuel in its reactors. “Those discussions are part of an on-going process,” according to Areva spokeswoman Cousinea.
The Energy Department has been counting on the government-owned Tennessee Valley Authority to use some of the MOX fuel, for example. But TVA spokesman Ray Golden said “we have made it clear that we have made no decision on whether we will use MOX or not.” He added that nothing was likely to change for “quite some time,” because the NRC still has not licensed the use of MOX fuel. TVA is presently having financial problems of its own and the Obama administration is considering selling it.
The Government Accountability Office says DOE has disclosed its intention to sell MOX fuel at 15 to 25 percent less than the cost of conventional fuel, but still has no takers. And if it did sell the fuel, the Congressional Research Service said that the government would likely see a return of just $1.1 billion to $2 billion on its investment of more than $20 billion.