By Jake Rossen
By Jesse Marx
By Michelle LeBow
By Alleen Brown
By Maggie LaMaack
By CP Staff
By Jesse Marx
According to the records, the metal tubing gave way at 2:13 p.m. A gash two inches long and half an inch wide burst open along a pipe carrying highly pressurized radioactive water, which proceeded to spew forth at close to 400 gallons a minute. At 2:33, someone picked up the "Red Phone" to notify the Nuclear Regulatory Commission of an accident at the Prairie Island nuclear plant.
Outside, things remained quiet with the exception of the cars streaming down the road as employees left the premises. Word that something might be wrong at the plant spread slowly from a bar until the rumor reached a radio station. The governor held a news conference at 5:30, announcing he had no idea what was going on.
The date was October 2, 1979, seven months after the near-catastrophe at Three Mile Island and seven years before the meltdown at Chernobyl. The incident at Prairie Island was comparatively trifling--a single rupture in the tubing of the plant's steam generator, which turns heat from the core into steam to power turbines. "All safeguards equipment functioned as required," Northern States Power told the NRC. Radioactivity was released into the air and the Mississippi River, but not at doses the government considered dangerous. Prairie Island was back on line in less than a month.
But almost 17 years later, the steam generator trouble haunts Prairie Island again. In April NSP Chairman and CEO James Howard stepped before shareholders at the company's annual meeting to announce that the utility had reached a settlement in a lawsuit against Westinghouse, which made the plant's four generators. Details about the case, and 16 similar suits nationwide, have been kept under wraps by the efforts of plaintiffs and defendants alike. Some of the documents that have emerged from the thicket of litigation indicate that nuclear plants around the country and the world may be sitting on a time bomb.
According to the papers--many of them never before made public--steam generators at dozens of nuclear plants, including Prairie Island, are deteriorating at an alarming pace. NSP alone has found thousands of cracks in the tubes that connect the reactor's smoldering core to the rest of the plant; some may be leaking radioactivity, and a few have burst. The utility has spent $62 million on repairs, and it has seriously considered replacing the equipment at a cost of at least $200 million more.
The story is the same at dozens of other nuclear plants nationwide, and it's getting worse so fast regulators can't keep up. At a recent NRC meeting, commission Executive Director James Taylor called steam generator degradation "one of the most serious challenges facing the industry today." His staff, he added, hadn't gotten around to writing new rules because they were busy "putting out fires." In a preliminary report presented at the same meeting, NRC researchers warned that under certain accident conditions, "significant radiological releases may occur." That's industryspeak for scenarios ranging all the way from a small radioactive plume over the neighborhood to a meltdown.
Even without a disaster--a possibility which, critics admit, is hard to pin down--the steam generator problem is rich with trouble for the industry, with a particular irony for NSP. "You'll remember how they told us [during the nuclear-waste fight in the 1994 legislative session] that Prairie Island is the most wonderful plant in the world," says Ken Tilsen, who's fought a local environmental group's court battle to win access to the documents. "Except that [they claim in the lawsuit] that the product is defective and they've been defrauded, lied to, and cheated over 25 to 30 years. And at the same time they tell the public that they've got the greatest plant that was ever built."
Bruce Drew has looked through 30,000 pages of documents over the last year or so, and that's only about half of the stash he's accrued. Steam generator papers have taken over the dining room of his south Minneapolis house--crowding hutches, spilling off credenzas, piled on every chair and tabletop. Drew worked as a chemical engineer and statistician until he retired a few years back. Now the NSP discovery project has become a full-time job. It's taught him, he says, more than he wanted to know about the innards of a nuclear plant--and, along the way, about what it takes to pry internal information out of a power company.
NSP filed its suit against Westinghouse in federal district court in St. Paul in the fall of 1993; close on its heels was the antinuclear Prairie Island Coalition, demanding "intervenor" status in the case--in other words, access to the documents the plaintiff and defendant were exchanging in discovery. There have been such requests in other such lawsuits around the country, and all have been rebuffed. But in mid-1994, U.S. Magistrate Franklin Noel granted the coalition's request.
When Drew and three other coalition representatives showed up seven months later for their first appointment at Prairie Island, they were told that half their group was prohibited from setting foot in the plant. The matter went back to court, and eventually a bizarre ritual evolved. Each day for more than a year, coalition members would show up in a special room at the downtown St. Paul offices of NSP's law firm. Boxes of documents would be brought in, which they'd peruse and mark for copying; often, NSP then declared the papers confidential. The coalition went back to court five more times, and the matter soon drew attention in high places. The giants of the international nuclear industry--France's Framatome, Germany's Siemens, and Canada's Babcock & Wilcox--all sent representatives to St. Paul to argue that the documents should be sealed. Noel ruled against all of them.
For the industry, steam generators are no marginal matter: They are, after all, at the very core of the process creating electricity out of fissioning atoms. In a pressurized-water reactor like Prairie Island, the radioactive core heats its coolant to 590 degrees; that liquid then runs through more than 6,000 horseshoe-shaped metal pipes, each about 70 feet long and almost an inch in diameter. When you run water along the outside of those pipes it immediately flashes to steam, which is piped to turbines. In a crude analogy, you could imagine the whole thing as a radiator, hooked up to a furnace, sitting in a bathtub. If the radiator breaks, the liquid in it mingles with the bathtub water; if a lot of fluid leaks out, the furnace could overheat.
To keep that from happening, Westinghouse pipes are manufactured from an alloy called Inconel-600, specially designed to the withstand punishing conditions of heat, pressure, and radioactivity. The problem, NSP's and other lawsuits allege, is that it doesn't. Though steam generators are supposed to last 40 years, tubes have been found to start cracking after less than 10 years of service. Engineers don't quite understand how this happens, but they know that the tiny fissures can grow larger and eventually burst.
And it turns out that Westinghouse may have been aware of this for some time--before, in fact, it sold most of its steam generators, including NSP's. (Westinghouse isn't the only company whose steam generators have been giving utilities trouble, but it's the only one so far being sued.) One 1964 Westinghouse memo--quoted in a lawsuit filed by South Carolina's Duke Power Co. in 1990--referred to an employee being told that "he was not to inform anyone with the exception of his boss of the Inconel corrosion problem, to prevent a possible hold on steam generator production." Another company document, dated 1968, showed a researcher's hand-written notation: "What do we tell them at this stage? That [Inconel] is crumbling before our eyes or that service experience is so far good?"
Judging from NSP's records, the former is closer to the truth. A 1991 memo from one of its researchers to the utility-funded Electric Power Institute bluntly noted that steam generators as designed were "a corrosion machine," and "no better than almost good enough" (emphasis in the original). By 1994, NSP had had to shut down its reactors at least eight times due to generator problems. More than 12 percent of a total of 13,500 tubes showed indications of cracking as of this February--enough to force NSP to keep its reactor shut down for two weeks beyond schedule for repairs.
"Repairs," of course, assumes that you know what you're fixing--which, in the case of steam generator troubles, reactor operators don't really. The tubes are inspected by a magnetic probe that travels through them on a cable. More sophisticated probes have been developed over the years, but documents from the lawsuit indicate that utilities aren't always eager to use the latest technology. In a handwritten 1986 note, NSP steam generator expert Richard Pearson summarized discussions with other company officials over a new kind of magnetic probe. It had been suggested, he wrote, that "we should tread very careful and not look for trouble... We don't have freedom to use the probe to look for as much information as possible."
Publicly, power companies have said that they "tread careful" with new testing equipment simply because the gear hasn't had all its kinks worked out yet. But it's also worth noting that anytime a utility has to plug or sleeve a steam generator tube, it loses a little of its power-generating capacity--and, by extension, revenue. "There has been a concern for some time," says Paul Gunter, who follows the issue at the Washington-based Nuclear Information and Resource Service, "that the industry is reluctant to have a full diagnostic done because they're worried about what they might find."
What's been found is enough to worry the Nuclear Regulatory Commission. Bill Russell, director of the agency's Office of Nuclear Reactor Regulation, told members at a recent meeting that "many more cracking indications than anticipated" were being found at plants throughout the country, and that their number was growing exponentially. Cracks would "jump from the tens in one [inspection], to the hundreds--or thousands--in the next." Even the materials used to fix the cracking tubes had been found to deteriorate in turn, forcing utilities to regularly redo earlier repairs.
The NRC and the utilities insist that none of this poses a danger to the public. "Our primary concern," says Brian Sheron, who heads the nuclear regulation office's engineering division, "is that these plants continue to operate with the tubes structurally intact. If that means that a plant has to shut down every five months, or if it means that a plant has to [fix] lots of tubes, then that may not make the utility happy, but it makes the NRC happy."
The utilities, in turn, say nothing makes them happier than going the extra mile for safety--particularly NSP, which now operates one of the oldest steam generators in the industry. Company officials say they've been able to keep the equipment going because of extra-thorough research and maintenance. (Prairie Island is frequently cited as one of the best-run plants nationwide.) "It's become just part of the way we do business," says plant manager Mike Wadley. "It consumes a lot of time and money and anxiety. But it's a manageable situation. And besides, we have defense in depth."
"Defense in depth"--the premise that each containment system is backed up by another containment system, which has its own backup, and so forth--is the principle at the heart of the nuclear industry's safety argument. Coincidence will have it, however, that the steam generator is one place where that rule doesn't quite apply. To protect the costly machine from overheating, engineers built into it a special vent designed to pop under pressure and release steam to the outside. When a tube breaks, that means highly radioactive steam goes out the roof--which is exactly what happened in 1979 at Prairie Island.
And 1979 was no fluke. In 1992, the NRC's then-director of nuclear reactor regulation, Thomas Murley, wrote in an internal memo that "there is a general acceptance that steam generator tube rupture events appear to be unavoidable." (Having just visited several reactors in Europe, Murley also noted that "it was clear that the U.S. lags behind the major European countries in terms of the scope of inspection.")
No cases of steam generator tube ruptures leading to catastrophic accidents are known, though the Mihama plant in Japan--whose generator was not made by Westinghouse--came awfully close in 1991. There, water rushed through a broken tube so fast the plant had to activate its emergency cooling system; some of that system's valves malfunctioned too, and technicians scrambled for some 45 minutes to arrest what could have been a meltdown.
And only a single tube broke at Mihama, an accident the steam generator is built to withstand. Multiple ruptures, by contrast, represent a "beyond design basis" condition, one where all bets are off. Though no one has wanted to contemplate the effects of such an incident, NRC Commissioner Kenneth Rogers told an industry symposium in 1988 that the scenario was entirely conceivable: If many tubes were deteriorating, as they were, it stood to reason that several could break at the same time. "In essence," he said, "we have a loaded gun, an accident waiting to happen."
So far, the NRC has taken a wait-and-see attitude to all this--occasionally asking for stepped-up inspections, yet allowing plants with severely degraded generators to continue operating. Last June, the Prairie Island Coalition asked the NRC to suspend NSP's license until the utility had inspected all of its tubes with state-of-the-art equipment. Russell notified the group that the petition "has not identified any safety issues warranting immediate action at the Prairie Island Nuclear Generating Plant... As provided by [federal law], NRC will take appropriate action on your Petition within a reasonable time." The group is still waiting.
NRC regulations, like any other kind, are not about safety at any cost. If the possibility of a particular kind of accident is deemed tolerably remote, or if contamination of people and the environment is kept within a certain limit, companies can't be forced to go to huge lengths to avert it. Thus, for example, there isn't much noise in the industry over the possibility that some of the cracked steam generator tubes may be leaking--which the documents indicate they do--because the amount of liquid is small enough to stay under a permitted "administrative leak rate" that works out to a gallon per minute at Prairie Island. (Interestingly, the 1992 memo from the NRC's Thomas Murley noted that limits on such releases were far stricter in Europe than in the U.S.)
And even the rules that govern when a utility needs to worry about steam generator cracks are up for changes. Utilities have long been telling the NRC that regulations written in the 1960s and '70s, for a whole different set of technical problems, are forcing them to go through unreasonable amounts of repairs and down time. "They are looking for some relief [from] a very conservative rule," says the commission's Sheron. "We've been dealing with that on a case-by-case basis, so we thought maybe we should promulgate a new rule." Sheron won't say what the new specifications will entail, and insists that "no one's getting a free ride here." But he knows, like any other federal employee, that the Clinton administration and Congress want agencies to "reengineer" themselves in a more customer-friendly manner. At the NRC, this has resulted in a promise of "cost-beneficial licensing actions," which translates as getting rid of particularly expensive requirement. In its 1994 annual report, the commission came right out and announced that "the industry [is] in the best position to identify which regulations impose a heavy economic burden with little commensurate safety benefit."
The phrase "economic burden" rings especially loud where nuclear plants are concerned because, as utilities have discovered over the past 20 years, the power once touted as being "too cheap to meter" is nothing of the sort. No new reactors have been ordered in the U.S. since 1978, and according to a 1995 study by the consulting firm Washington International Energy Group, three-quarters of utility executives surveyed didn't think their companies would ever build another nuclear plant. Even Hazel O'Leary, the former NSP exec who now heads the Department of Energy, told Congress in 1993 that "the costs of nuclear power, if you include new construction, are not competitive." Against that backdrop, the extra cost of deteriorating steam generators (and other reactor parts) may be the last straw for some plants. At least one utility, Oregon's Portland General Electric, has already decided to just chuck the whole thing. For years, citizen groups demanded the shutdown of the Trojan reactor, and PGE resisted. In 1992, it turned around and closed the plant--in large part because of the escalating cost of steam generator repairs.
At NSP--whose CEO, Howard, once said he didn't think nuclear power had an economic future--things haven't reached that point: Company executives insist that Prairie Island still provides some of the cheapest power in its system. But the steam generator price tag isn't insignificant either: Pearson calculates that the company spends about $2 million every 18 months on repairs and inspections. And NSP recently picked up a whole new set of cracking generators when it merged with Wisconsin Electric Power Co., operator of the Point Beach nuclear plant. The utility is allowed to recoup most of the money for repairs through increases in electricity prices; the company says it will pass the savings from the settlement with Westinghouse on to consumers, but since precise figures aren't public that claim can't be verified.
One theory holds that Westinghouse has been willing to settle with power companies--none of the 17 steam generator suits against it has gone to trial--because it must keep open one of the few markets left for its nuclear division. As prospects for new reactor sales in the U.S. grow dim, manufacturers are left with only two options: selling abroad (a few Asian countries, notably China, have aggressive nuclear energy programs), and selling replacement parts such as new steam generators. As a Westinghouse spokesman offered in the company's sole pronouncement on the NSP settlement, the utility was "a valued customer, and we prefer not to fight with valued customers in court."
Thus, while Westinghouse and the power companies may be at each other's throats in court, they share a fundamental interest--diverting public scrutiny of the steam generator problem. "This litigation is harmful to utilities," a Westinghouse executive warned in a 1993 letter to the Tennessee Valley Authority. "For example, the Union of Concerned Scientists has used the litigation as a vehicle to incorrectly imply that steam generator issues pose health and safety risks to the public. This message has been communicated to the media and to legislators serving in the U.S. House of Representatives and the Senate.
"If the current litigation process proceeds through the public trial stage, we will have created a platform for those opposed to nuclear power to unfairly attack both the safety and economics of operating nuclear power plants. The public spectacle that steam generator trials will create will further threaten the nuclear power option for the future of our nation."
Right now, it doesn't look as if the public will get much of a spectacle out of the NSP case. The utility abruptly and unilaterally terminated the Prairie Island Coalition's discovery in January, saying the allotted time had run out. Magistrate Noel rejected that idea and slapped the utility with a $2,000 fine. Yet by March, NSP won a ruling from U.S. District Court Judge James Rosenbaum that even the possibility of a settlement made further discovery "moot." The coalition, he said, couldn't even get documents NSP had been ordered to release almost a year before. The group is fighting that ruling in the Eighth District Court of Appeals.
Even without the 14,000-or-so pages it's still trying to get, the coalition has its work cut out. Most of its 60,000-page cache has yet to be reviewed in any detail; a troupe of volunteers expect to be at it for at least the rest of the year.
It's hard to say what they'll find. Many of the documents reviewed so far are highly technical and routine in nature, day-to-day notations of industrial practice. Two years in, Drew says he's only beginning to see the outlines of a bigger picture. "The overwhelming sense you get is of a huge industry dealing with a very dangerous technology which they're working very hard to keep from going awry--and at the same time working very hard to keep the public from knowing that they have to work so hard to keep it from going awry. And I wonder when it's going to catch up with them. I think it'll be during my lifetime." He's 72.