09/22/02

Washington

- Yielding to lobbying by the nuclear power industry and pressure from Congress, the Nuclear Regulatory Commission has changed - some would say relaxed - the way it regulates in recent years.

The change has saved the companies millions of dollars in compliance costs, but in some cases it also has risked safety, according to interviews with industry and regulatory insiders and critics and an extensive review of public documents. A Plain Dealer review shows:

The nuclear power industry, a savvy political player with deep pockets, leaned on Congress when it felt that NRC inspectors were being too harsh and wouldn't listen to its pleas. Congress in turn pressured the NRC, threatening to slice the agency's budget unless it backed off.

The nuclear industry profited from the changes - the Nuclear Energy Institute, the industry's lobbying group, estimated from a 1999 survey that each of the nation's 104 plants could save as much as $5 million a year with the regulatory changes.

Those changes altered the NRC's focus, shifting its sweeping enforcement strategy to one centered on areas most critical to safety. But the new focus relies heavily on self-policing by the corporations that own the power plants and on assumptions about what could go wrong. Those assumptions haven't always proved correct.

The worst known case of false assumptions was at the Davis-Besse nuclear power plant in Oak Harbor, near Toledo. Stainless-steel nozzles that pass through the lid above the reactor cracked, leaking boric acid - which ate an 8-inch-diameter hole in the lid. Only a thin stainless-steel liner kept the reactor's high-pressure coolant from causing a serious accident.

The hole, discovered in March, was the result of years of neglect by plant owner FirstEnergy Corp. of Akron - and the industrywide assumption that a failure of the lid would be unlikely.

The plant, meanwhile, had received the NRC's top safety ratings.

Earlier error

The NRC's failure was reminiscent of an embarrassment in February 2000 when a steam generator tube ruptured at Consolidated Edison's Indian Point nuclear power plant north of New York City. A small amount of radioactive material leaked into the Hudson River.

As at Davis-Besse, NRC staff members had concerns that some plant components at the Indian Point plant might be weak. But the NRC didn't pay close attention to the methodology used in an engineering report from the plant that said everything was fine. It turned out that the methodology was badly flawed. Instead of questioning the methodology, the NRC took the utility at its word, the agency's inspector general later reported. The reasons: The NRC project manager assigned to the plant said that steam generators were outside his area of expertise and that the staff generally relied on the utilities to evaluate their own data.

Likewise, a junior NRC engineer who had concerns about Indian Point was afraid to ask follow-up questions - because, according to the inspector general's investigation, she was told not to burden the utility unnecessarily. The agency "frowned upon" asking a utility more than one round of questions, she reported.

This is the rap the NRC faces: It is "too close to the nuclear power industry." So says U.S. Sen. Harry Reid of Nevada, who has dealt with the NRC for years and is bitter about losing the fight to keep nuclear waste out of his state. "They are a lapdog for the nuclear power industry," he says.

He is not the first to make that accusation. Fifteen years ago, the House Interior Committee issued a 44-page investigative report with a title that neatly summed up its findings: "NRC Coziness With Industry."

NRC Chairman Richard Meserve calls the criticism "very unfair." He and others at the agency say nuclear plants are safer than ever, a result of the decades of experience that power plant operators and the government now have. The agency cites statistics that show fewer workers are exposed to radiation, fewer shutdowns are required for emergencies, and more plants are running more hours and generating more electricity than ever.

"If you look at all these indicators, you've got this slope that comes down to everything's working today," says Frank Gillespie, an NRC deputy director in charge of regulatory improvement.

Numbers game

Some of these statistics are skewed, note critics such as the environmental groups Greenpeace and Public Citizen - plants don't shut down as often in part because the NRC now lets them keep operating while they make certain repairs, unlike in the past. But most skeptics don't argue with the overall record.

"By almost any measure, Davis-Besse being a horrible exception, the performance of the industry in terms of objective indicators has improved pretty substantially over the past decade," says Edward McGaffigan Jr., one of the NRC's five commissioners.

But it would take only one bad accident to turn those statistics on their ear. And the NRC's more lenient approach during the last two years toward oversight leaves too big a risk that a nuclear accident will occur, critics say. The agency, formed to protect public safety by keeping a sharp eye on the industry, instead has become a cheerleader for nuclear power, say authorities who include former NRC commissioners.

With Davis-Besse, the NRC "seemed to regard its mission as putting out a lot of reassurances that were at odds with the actual evidence," says Peter Bradford, a former NRC commissioner.

"There was this attitude of patting people on the back and saying, 'OK, go back to work,' while they really didn't have a clue," says Victor Gilinsky, another former commissioner.

The NRC operates on a set of engineering-backed assumptions that help it focus on power plants, systems and equipment that need more thorough and frequent monitoring. The system is "risk-informed" in NRC parlance: Engineers use complex formulas and matrices based on experience and studies to decide which components have the highest risk of breaking or leaking.

The system is designed to help nuclear plant operators and inspectors prioritize. It lets an inspector "adjust how often you look at things that are running well and try to focus your inspection on things that are not running well," says Gillespie, who helped shepherd in this new regulatory focus.

System upgrade

The agency started phasing in the system in 1999 and fully implemented it in 2000. NRC officials say that they had wanted an analytical system for years because it would offer a more objective approach but that they had to wait for the necessary computer modeling to improve.

The nuclear industry and the Nuclear Energy Institute, its lobbying arm, had been pushing for risk-informed oversight as well. It complained that the old regulatory system - a prescriptive series of rules enforced unevenly - did not distinguish between risky violations and benign infractions.

That meant citations for having an electrical equipment operating manual that failed to include an instruction to plug the equipment in, says Steve Floyd, NEI senior director of regulatory reform, citing real examples. Another plant was written up after an employee left a book atop a control panel. As the inspector reasoned, a minor earthquake could have shaken the panel and the book could then have hit a control button.

NRC inspectors, says Floyd, were "running amok" looking for infractions.

The NEI was so interested in risk-informed oversight that it prepared a policy paper and held a conference in Orlando, Fla., in 1996 to talk it over with nuclear plant operators. But on the morning those executives were on their way to the airport for the conference, a bombshell dropped that set their efforts back several years: Time magazine came out with a cover story on how poorly the NRC was regulating the plants.

The Time story focused on the Millstone nuclear plant in Waterford, Conn., which had been improperly handling spent nuclear fuel and ignoring safety requirements. When a conscientious worker complained, the plant's owner ignored him. In despair, he went to the NRC - and found that the agency had known about the problem for a decade but never tried to stop it.

The NRC, said Time, "may be more concerned with propping up an embattled, economically straitened industry than with ensuring public safety." And the Millstone plant, it said, "is merely the latest in a long string of cases in which the NRC bungled its mandate and overlooked serious safety problems until whistleblowers came forward."

Indeed, the General Accounting Office, the investigative arm of Congress, found that other utilities, too, were not correcting their safety deficiencies - and that the NRC allowed the problems to persist.

Temporary fix

The result was predictable: Congress held hearings, a chastened NRC swore it would improve, and the NEI shelved its proposal for regulatory change.

By 1998, the power plants were chafing under the NRC's leash. The NEI went to Congress, complaining to sympathetic members about the NRC - not that the agency was tough but that it was petty, inconsistent and unpredictable.

"We had a lot of problems with the NRC, and we realized we were not going to get the NRC's attention unless we went to their boss," recalls Floyd. "And Congress is their boss."

One longtime nuclear-power supporter, Sen. Pete Domenici, a New Mexico Republican, chaired a Senate appropriations subcommittee that oversaw the NRC's budget. In a hearing in spring 1998, he told the agency to back off - or he'd cut its budget for plant inspections by 40 percent.

The NRC heard him loud and clear. "Shirley Jackson is a smart lady," says Floyd, referring to the woman then in charge of the NRC, "and she got the message from the congressional hearing."

Jackson, now president of Rensselaer Polytechnic Institute in upstate New York, did not respond to several requests for an interview. But Commissioner McGaffigan says of Domenici: "We needed a kick in the pants back when he gave us the kick in the pants."

Working with the Nuclear Energy Institute, the agency wrote up its new plans. Goaded by Congress, it also fine-tuned several strategic goals, among them: Don't burden power plants with unnecessary regulations.

"In all honesty, this agency goes through cycles," McGaffigan says. "After Three Mile Island [in 1979], people tell me that we just put a whole bunch of rules on the books. And when you look at them from the point of view of their cost-benefit effectiveness, they don't hack it." Changing the focus did more than save money, he says; it helped the agency prioritize its inspections and enforcement.

Fuzzy math

What worries skeptics is the mathematically modeled assumptions used in deciding where to put those priorities. For instance, the NRC and the industry decided that some components - such as reactor vessel lids - posed little risk. After all, how could a stationary piece of steel 6½ inches thick possibly break?

But the industry had been aware of the cracking of stainless steel nozzles in the lid - the source of the Davis-Besse corrosion, it later turned out - for more than a decade. In France, the state-owned utility in the early 1990s began a program of replacing vessel lids on all its reactors because of concern about such cracking.

To the NRC, however, the "chances of failure [of the lid] were so low that it didn't even appear on their radar screen," says David Lochbaum, a nuclear-safety engineer with the Union of Concerned Scientists.

As a result, keeping an eye on the lid was a low priority to the NRC's two resident inspectors at Davis-Besse. The agency sent out periodic notices warning that nozzles in contact with boric acid could be prone to cracking and that boric acid was corrosive. But it was up to the plants, not the NRC, to check for damage. The nozzles allow rods that control the nuclear reaction to pass through the lid.

Asked how the resident inspectors could have overlooked clues to the problem - including the repeated clogging of air filters with a brownish substance that proved to be airborne rust from the corroding lid - NRC officials said the inspectors had many other matters to attend to.

In that regard, the steel lid on Davis-Besse may represent the biggest flaw in the NRC's new oversight system: Its assumptions may or may not be correct.

"There are still unknowns after all these years," says Thomas Murley, a retired NRC director of reactor regulation. "That was not foreseen, and therefore it was not modeled in the risk assessment."

Gillespie, the deputy director who defends the system, allows this: "There is an old quote that somebody in my 30 years of history said: 'The industry and the NRC are very good at preventing what we know about. But it's the instance that we haven't preplanned for that's going to get us.' "

Overlooked area

Some within the NRC suggest that too many such instances are possible. For one, many of the plant analyses do not adequately take into account the risk of human error, according to a recent internal NRC draft report.

George Lanik, a team leader in the NRC branch that studies regulatory effectiveness, examined problems at power plants from 1993 through 2000. He looked at the "accident precursors" - events that led to the problems in the first place - and many were caused by human error.

Some 42 percent of the incidents in what could have become serious accidents were "due to the events not typically modeled" on risk assessment, Lanik wrote.

"It's the surprises, and the things that we haven't got in those models yet, that are going to come back to bite us," he says.

But to avoid being unpleasantly surprised, top NRC officials say, every nuclear power plant has a series of safety systems that are intended to preclude or contain virtually any disaster before it hurts the public. NRC officials noted until recently that, despite the corrosion in the Davis-Besse lid, the stainless steel liner on the underside of the damaged head kept the radioactive coolant contained. But the NRC officials learned this month that that notion, too, might give dangerously false comfort. Subsequent tests showed that the liner was cracked and thinner than they had known. Further, they acknowledge, the liner was never intended to be a part of the safety system.

The NRC is reviewing just what went wrong at Davis-Besse and how much of the blame it shares. An NRC task force expects to release a "lessons learned" report by the end of the month. NRC criminal investigators have been looking into why NRC officials allowed Davis-Besse to keep operating beyond the end of December when they knew there could be a problem.

The agency is clearly expecting to wind up with egg on its face. Says Chairman Meserve: "I don't know what the lessons-learned task force at Davis-Besse is going to say. But I presume that they will have some comments about whether there were failures in our oversight program that need to be addressed.

"I expect that would be fertile ground with them," Meserve says. "And if there are, this is something we're going to change."

To reach this Plain Dealer reporter:

skoff@plaind.com, 216-999-4212