Advertisement

How San Onofre’s new steam generators sealed nuclear plant’s fate

Aerial view of the San Onofre Nuclear Generating Station in the northwestern corner of San Diego County. Problems with replacement steam generators led to the plant's permanent shutdown.
(Don Kelsen / Los Angeles Times)
Share

In March 2004, an attorney for Southern California Edison sat before state utility regulators to propose what seemed like a great deal.

The San Onofre nuclear plant was approaching the end of its life span. But Edison wanted to invest $680 million in new steam generators, attorney Carol Schmid-Frazee told a judge presiding over a hearing at the California Public Utilities Commission’s San Francisco headquarters. The new equipment, she said, would give the 2,200-megawatt plant a new lease on life, providing cheap, reliable energy in Southern California for decades to come while also saving ratepayers nearly $2 billion.

Edison’s lawyer also issued an ominous warning: If regulators did not approve the upgrade, the plant would close, provoking “very serious problems with the California electric grid.”

Advertisement

The commission was persuaded, and Edison began remaking San Onofre.

But less than a year after the new steam generators came online, a tube in one of them sprang a radioactive leak, setting off a chain of events that ultimately led Edison to close the plant permanently. The generators that were supposed to save San Onofre ended up killing it, and today the atomic behemoth sits idle, never again to produce a watt of power.

Instead of a boon to California, the steam generator replacement became a billion-dollar debacle, leaving 1,100 workers facing layoff and the plant’s owners, suppliers, regulators and customers wrangling over who is to blame, how to replace the plant’s power — and who will pay for the mess.

::

The affair began with good intentions.

The nuclear plant, one of only two in the state, had been an eye-catching fixture on San Diego County’s north coast since 1968, powering about 1.4 million homes. But by 2001, one of its reactors had permanently shut down and the other two were showing their age. Edison believed modernizing San Onofre would provide cheaper electricity than building new natural gas plants or purchasing imported power.

That year, Edison began a formal study of the plant’s biggest problem: cracking and wear on the tubes in its original steam generators, which produce power and help cool the reactor core. Thousands of narrow tubes carry superheated water from the reactor that is used to create the steam that turns electricity-producing turbines.

The four generators, each as long as an 18-wheeler, were supposed to last through the end of San Onofre’s operating license in 2022. But, like most steam generators of the era, their tubes had corroded, and Edison feared it might have to close the plant more than a decade early if it didn’t replace them.

Advertisement

Doing so was expected to aid its chances of getting a 20-year license renewal from the U.S. Nuclear Regulatory Commission. It was a big job, but a routine one: By 2004, nearly three dozen other plants in the U.S. had already swapped out their steam generators.

Not everyone was on board. The plant’s minority co-owners — San Diego Gas & Electric and the cities of Anaheim and Riverside — protested that Edison had embarked on the project without their input. Filings with the PUC spoke of “acrimony” between the parties.

SDG&E; President Mike Niggli said his company “had serious questions about whether it would be the right investment.”

Anaheim pulled out, dropping its 3% ownership stake. SDG&E;, which initially refused to participate, finally agreed to contribute 20% of the cost, and Riverside — which owns less than 2% of the plant — also went along.

In September 2004, Edison selected Japanese company Mitsubishi Heavy Industries to design and build the generators. Mitsubishi had made more than 30 nuclear steam generators in other countries and was replacing the ones at Nebraska’s Fort Calhoun plant, which were similar to San Onofre’s.

In fact, though, Mitsubishi’s engineers had never made steam generators as large as San Onofre’s, each of which stood more than 65 feet tall, weighed 620 tons and contained more than 9,000 tubes.

Advertisement

From the outset, Edison was worried about the design of the generators’ anti-vibration bars — support structures that are supposed to prevent the tubes from shaking.

“I am concerned that there is the potential that design flaws could be inadvertently introduced” in the large generators, wrote Edison’s then-vice president in a November 2004 letter to Mitsubishi. He worried, he added, about “unacceptable consequences.”

Edison and Mitsubishi formed a special committee to deal with the issue, which met for more than a year to debate possible solutions.

The final design incorporated significant changes. The companies swapped out the alloy used in the tubes for one less susceptible to corrosion. Because the new alloy conducts less heat, they also added 377 more tubes in each steam generator. Other plants, including Fort Calhoun, had made similar changes and experienced no problems.

Problems surfaced repeatedly during manufacturing, leading to substantial delays, but the new generators were finally installed and the reactors began producing power.

::

Advertisement

At first, the upgrade appeared to be a success.

“This is a moment of pride,” Edison chief nuclear officer Pete Dietrich said in February 2011, adding that the project “ensures that the plant’s contribution to Southern California will continue for another decade or more.”

The next January, the plant’s Unit 2 reactor was powered down for refueling. Routine inspections of its steam generators found 1,595 tubes with wear, mostly from rubbing against support structures. That was unusual, but most of the damage was minor and Edison chalked it up to “settling in.”

Workers made repairs and were preparing to bring Unit 2 back online, when at 3:05 p.m. on Jan. 31, an alarm sounded in the control room. A tube in one of the 11-month-old steam generators in Unit 3 had sprung a pinhole leak.

The amount of radioactive steam spewing out — about 75 gallons a day — was small enough that the unit could have legally continued running, but operators immediately powered the reactor down. On-site inspectors from the NRC hurried to the control room, and calls and emails flew back and forth between Edison and Mitsubishi officials.

“I knew immediately it was going to be an issue,” said Frank Gillespie, a top executive in Mitsubishi’s nuclear division. “How big an issue it was, I didn’t really have a sense for.”

With San Onofre dark, engineers began an intensive series of tests.

The findings were troubling: 1,806 tubes in Unit 3 showed wear. What’s more, it was considerably more severe than in Unit 2, and some was of a type never before seen, caused by tubes knocking against each other. Eight failed pressure tests.

Advertisement

Alarmed, a group of five high-level Edison officials, including Chief Executive Ted Craver and Dietrich, called a series of closed-door meetings. Keen to minimize losses, they discussed restarting Unit 2, but Craver argued they should keep both reactors off until they had figured out what went wrong.

“No amount of financial stress or anything else would equal what we would go through if we brought the unit back online and had a problem,” he said in an interview. “If it’s a real disaster, if it’s a Fukushima kind of thing, that could be the end of the company.”

Eventually, all five men agreed to keep Unit 2 offline, and on March 27, 2012, the NRC sent a letter ordering both reactors shut down until Edison found a solution. Meanwhile, the NRC launched its own investigation.

It quickly became obvious that the leak, which engineers assumed was an isolated issue caused by a loose piece of metal rattling around, was part of a much bigger problem. An extended outage or major repairs would be costly; everyone began looking for someone to blame.

The NRC delivered its initial conclusions in June, blaming faulty computer modeling by Mitsubishi. Because of a single incorrect measurement in the code, computers vastly underestimated how fast steam would be flowing.

High-velocity steam in turn had caused the tubes to vibrate, banging into supports and other tubes. Because of slight differences in the manufacturing of the support structures, it appeared that Unit 3 was more prone to such vibration than Unit 2, which could explain why it had so much more wear.

Advertisement

Mitsubishi maintains that correct computer modeling wouldn’t have made a difference, because engineers were focused on preventing a different type of vibration that had been a problem at other plants.

By last October, the outage had cost Edison more than $300 million, and the longer San Onofre remained offline, the more likely it was that shareholders would end up with the bill. California utilities can pass on a plant’s costs to ratepayers, but if it does not produce power for nine months, the PUC can force the company to issue refunds.

That month, Edison went to the NRC with an unorthodox proposal to restart Unit 2 and run it at 70% power on the theory that reducing the power would reduce vibration. After five months, the unit would be shut down again for inspections, and if there was no more damage, it could keep running — perhaps at higher power.

The proposal didn’t mention the fate of Unit 3. Company officials were quietly discussing the possibility that it would never fire up again.

Meanwhile, tensions were rising between Edison and Mitsubishi. In a series of testy letters in late 2012, Dietrich chastised Mitsubishi for not having proposed a viable long-term fix. In response, Mitsubishi officials suggested installing thicker anti-vibration bars, which could be done in about a year. The alternative, replacing most or all of the steam generator system, would take more than five years.

::

Advertisement

As Edison struggled to come up with a restart plan, the future of San Onofre had also become a political flash point.

Critics had for years pointed to a history of safety violations at San Onofre, among them NRC citations for failed diesel generators, inoperable emergency power batteries and falsified fire safety records. After the 2011 disaster at Japan’s Fukushima Daiichi nuclear plant, communities around San Onofre grew nervous about the possibility that it, too, could be overwhelmed by an earthquake or tsunami.

Soon after the leak, environmental group Friends of the Earth waded into the fray with an influential ally — Sen. Barbara Boxer. The California Democrat chairs the Senate committee that oversees the NRC.

Boxer said she first grew concerned about San Onofre’s safety after the Fukushima meltdown. She and activists argued that Edison should have gotten additional approvals from the NRC before installing its new steam generators because their design was so different.

Edison insists it followed standard industry practice and that the generators were meant to be an improvement.

NRC staff concluded that Edison had followed regulations, but the issue gained traction after Boxer and Rep. Ed Markey (D-Mass.) released portions of a leaked Mitsubishi document they said showed that Edison knew the steam generators were flawed but installed them anyway.

Advertisement

Activists hailed Boxer as a hero. Craver complained about the injection of “politics into a process that should be free from it.”

Edison had hoped to have Unit 2 back online by this summer, but the NRC’s decision was repeatedly delayed. Then in May, a panel of NRC administrative judges ruled that Edison might have to go through a lengthy license amendment process before restart — the outcome that Boxer and the activists had been pushing for.

“At that point, I believed it was preferable to shut it down,” Boxer said.

Edison, meanwhile, was losing patience. Its analysts found that keeping the plant at the ready — at a cost of about $1 million a day — was unsustainable. Natural gas prices had plummeted, making replacement power cheaper, and the PUC still had not determined whether the costs of the outage would fall on customers or shareholders. If the plant did not restart before the end of 2013, the math got ugly.

To handicap its options, Edison hired a law firm to analyze the odds of regulatory approval, Craver said. The firm found a less than 50% chance of a resolution before year’s end.

On June 6, Edison’s board held a telephone conference. The vote was final. They were pulling the plug.

“It no longer makes sense for us to seek restart of San Onofre,” Craver told investors the next morning. Within hours, Boxer said in a statement she was “relieved” by the news.

Advertisement

With the plant’s fate settled, the next battle will be over money. The PUC will have to decide how to allocate a total bill for the debacle that is now approaching $2 billion — about the amount Edison once promised that the generators would save customers. Litigation between the companies is widely expected.

Dave Wright, general manager of Riverside Public Utilities, voiced the feelings of many involved in the ill-fated project:

“In hindsight, if we knew we were going to get no benefit out of this major investment, we never would have done it,” he said. “But this is one of those who-would-have-thunk-it moments. It was inconceivable.”

abby.sewell@latimes.com

ken.bensinger@latimes.com

Advertisement