Union of Concerned Scientists
programmatic breakdowns are unlikely to be con- fined to only one of the reactors at a given site.
And yet, companies operating multiple nucle- ar power plants seldom shut down all of their reac- tors when one plant experiences a programmatic breakdown. This makes little sense considering that programmatic breakdowns are unlikely to be confined to only one facility owned by a poorly managed company. The NRC must determine whether programmatic breakdowns identified at one site also affect reactors operated by the same company at other sites. The agency must also ensure that a plant owner’s focus on restarting a troubled reactor does not cause or contribute to declining performance at other reactors owned by that company.
Lesson 9: Better Communication Is Needed inside the NRC
The NRC quite properly does not accept excuses from a plant owner such as, “The maintenance department knew about an unresolved equipment problem but failed to inform the operations department.” As documented in our case studies, however, the NRC repeatedly suffers from inter- nal communication barriers of its own.
For example, NRC Region III staff knew about containment problems identified by its inspectors at Donald C. Cook, but was unaware of related problems that the staff at NRC head- quarters knew existed at other plants. This pat- tern recurred five years later when UCS had to inform Region III staff about problems known to NRC headquarters that were applicable to the regional team’s work at Davis-Besse. The NRC must develop more effective internal communica- tions so that knowledge possessed by one region or office is shared by all.
Lesson 10: Not All Poor Performers Have Had a Year-plus Outage
Though it was not an explicit objective of this analysis, we came across five reactors that were
not on the year-plus outage list even though their documented performance levels were as bad as, or worse, than many reactors on the list. Dresden Units 2 and 3 in Morris, IL, for example, were on the NRC’s Watch List for 7.5 years—longer than any U.S. reactor other than Browns Ferry—but never experienced a year-plus outage. Seventy-five percent of the reactors on the Watch List for a year or longer experienced a year-plus outage, and all of the reactors on the Watch List for three years or longer—except for Dresden—experienced a year-plus outage. Apparently, the NRC believed Dresden had the same performance problems as the other reactors but, for some reason, did not need the same remedy.
New Jersey’s Salem Units 1 and 2 experi- enced a year-plus outage in the mid-1990s, but they, along with the Hope Creek reactor literally next-door, did not meet a similar fate in 2003 despite ample evidence from several independent assessments that safety levels were at least as bad as they were at the time of the mid-1990s out- age. Salem had clearly suffered a relapse, but the NRC inexplicably opted for a different treatment regimen the second time around.
Lessons Must Translate into Action
The bad news about the many lessons to be learned from the 51 year-plus outages is that much work needs to be done by the NRC to more effectively oversee safety levels at nuclear power reactors. The good news is that the NRC, when working effectively, can produce successful outcomes. The challenge and imperative is thus to move this needed work from the road ahead to the rearview mirror.