San Onofre design choices led to nuclear plant shutdown
The company made choices in designing support structures at San Onofre that were intended to prevent one type of vibration, but ended up creating another type of vibration that ultimately led to the plant's closure, said Frank Gillespie, senior vice president with Mitsubishi Nuclear Energy Systems.
The problematic vibration, he said, had not been seen at any other plant before, although it had been observed in experimental conditions.
That vibration led to excessive wear on the tubes, particularly in the plant's Unit 3, where one tube sprang a leak and released a small amount of radioactive steam on Jan. 31, 2012, and eight tubes failed pressure tests.
The nuclear facility has been closed for more than a year.
Mitsubushi discussed the design process in a proprietary report that was made public in a redacted form earlier this month.
Gillespie said designers working on the new system in 2005 put "paramount focus" on controlling vibration and reducing wear. In the process, they added more anti-vibration bars, but made other changes that led to less contact between the bars and tubes.
In Unit 3 in particular, the bars were flatter, leading to about half the amount of pressure between bars and tubes as in Unit 2, the plant's other working reactor unit, which also saw an unusual but less severe amount of wear.
“What they didn’t understand at the time is, some of the steps ... actually made in plane [vibration] worse,” Gillespie said. "...There was an underappreciation for the fact that the pressure of the bars against the tubes actually performed a very important function."
Anti-nuclear activists and some lawmakers -- notably, Sen. Barbara Boxer (D-Calif.) and U.S. Rep. Ed Markey (D-Mass.) -- have accused Mitsubishi and plant operator Southern California Edison of being aware of defects in the equipment's design prior to installation and failing to make modifications that might have prevented the problem in order to avoid going through a potentially lengthy license amendment process.
Mitsubishi's root cause report did show that some changes were rejected in part because they would have required a license amendment. The changes were intended to reduce the dryness of the steam flowing around the tubes, which ended up being a factor in the problematic vibration.
Gillespie said the manufacturing contract required that the new steam generators should fit in the same physical space as the old ones. Some of the changes that were considered and rejected would have changed the size of the apparatus enough that it would not have fit, which would have required a license amendment.
But, he said, “We totally thought we had eliminated the right kind of vibration. There was never a decision to do something that would have made this generator, in 2005 considerations, less safe.”
Some changes in the design made to conform with changing industry standards made the problem worse. The alloy of the tubes was changed to prevent corrosion. Because the new alloy conducts less heat, that meant 10% more tubes were added. The tube design was also changed so that a bend at the top of the tube bundle was more rounded.
Gillespie said those changes were "standard procedure" in the industry, but they contributed to the conditions that led to the vibration. The design process was complicated by the fact that San Onofre's original steam generators are unusually large in physical size and power output.
Mitsubishi also made an input error in a computer code intended to predict the velocity of the steam, which led to underestimating how fast the steam would flow around the tubes.
A U.S. Nuclear Regulator Commission team tasked with investigating the issues at San Onofre pointed to that error as one of the major causes of the plant's issues, but Mitsubishi maintains that even if the calculation had been correct, they would not have predicted the in-plane vibration that occurred.
The NRC is weighing a proposal by Edison to restart Unit 2 and run it at reduced power for five months, in hopes that reducing the power would eliminate the conditions that led to the unwanted vibration. The agency will not make a decision on the plan until May at the earliest.
In the meantime, an NRC panel is weighing a petition by environmental group Friends of the Earth to require a license amendment process including trial-like public hearings, before deciding on the restart proposal.
Friends of the Earth has argued that Unit 2 is unsafe to run without major repairs or replacing the steam generators again.
Gillespie disagreed, saying he considered it "highly likely that Unit 2 is actually a very good steam generator" and that preventatively plugging tubes that might be susceptible to wear should address the problem.
No restart plan has been proposed for Unit 3. Gillespie said Mitsubishi had recommended several options to Edison for addressing the issues in that unit, but would not discuss what the options were in detail. Edison said in a financial filing that one of the recommendations, which would entail replacing substantial portions of the steam generator system, could take as much as five years.
The two companies are also at odds over a $138-million cap on the steam generator warranty, which Edison believes should not apply to the current situation.
-- Abby Sewell
Photo: Beach activity near the San Onofre nuclear plant. Credit: Allen J. Schaben / Los Angeles Times