Nuclear Safety Performance at Pilgrim

February 6, 2017 | 6:00 am
Dave Lochbaum
Former Contributor

The Nuclear Regulatory Commission (NRC) held a public meeting on Tuesday, January 31, 2017, in Plymouth, Massachusetts. A large crowd of over 300 individuals (perhaps thousands more by White House math) attended, including me. Elected officials in Massachusetts—the attorney general, the governor, the entire US Congressional delegation, and state senators and representatives—had requested the meeting. Many of these officials, or their representatives, attended the meeting.

The elected officials asked the NRC to conduct a public meeting to discuss the contents of an email from the leader of an NRC inspection team at Pilgrim to others within the agency regarding the results from the first week’s efforts. An NRC staffer forwarded this email to others within the agency, and inadvertently to Diane Turco of the Cape Downwinders, a local organization. The contents of the leaked email generated considerable attention.

Unique NRC Meeting
During my nearly two decades at UCS, I have attended dozens, perhaps hundreds (maybe even millions by White House accounting) of NRC meetings. The Plymouth meeting was unique. It was the only NRC meeting I’ve attended to discuss an email.

And it was the only NRC meeting I’ve attended where public speaking slots were chosen by raffle. In all prior meetings, members of the public raised their hands to be called upon by the NRC staff, queued behind a microphone in the room in order to speak, or added their names to a list to speak in the order specified by the sign-up sheet. At this meeting, the NRC used a raffle system. I received Ticket #4 (see Figure 1), giving me an opportunity to “win” a chance to speak for up to 3 minutes (or 180 seconds, whichever came first) during the meeting.

Fig. 1 (Source: Nuclear Regulatory Commission)

Fig. 2 (Source: Nuclear Regulatory Commission)

My ticket, along with at least 74 other tickets, was placed into a fishbowl. Brett Klukan, an attorney in NRC Region I, drew tickets from the bowl to establish the speaker order. Because the fishbowl was clear glass, Brett gazed at the ceiling to avoid charges of cherry-picking preferred ticket numbers (see Figure 2). Brett then wrote the number drawn on a whiteboard without showing the number to anyone else, somewhat offsetting the averted gaze tactic since he could have jotted down any number he wished.Unique NRC Discussion

Brett Klukan opened the meeting by introducing the NRC panelists and covering some ground rules for the meeting. The ground rules included a decorum standard—any audience member disrupting the meeting three times would be asked to leave. If the individual did not leave voluntarily, Brett explained that law enforcement officers (and there were numerous uniformed officers in the room and in the hallway outside) would escort the person from the room.

Brett then turned the meeting over to the NRC panel of Dan Dorman, the Regional Administrator for NRC’s Region I, Bill Dean, the NRC’s Director of the Office of Nuclear Reactor Regulation, Raymond Lorson, the Director of the Division of Reactor Safety in Region I, and Don Jackson, the leader of the NRC inspection team at Pilgrim and author of the email.

Don went through the leaked email, which he had written, updating the audience on each issue and supplementing the email with results from the team’s efforts since that initial week. I had expected the NRC to talk about what systems, components, and administrative processes the inspection team examined, but anticipated the NRC would not discuss results until the team’s report was approved and publicly released. But Don candidly provided the results, too. More than once, Don explained that the team identified an apparent violation of NRC’s regulations—in fact, he stated that 10 to 15 potential violations had been identified.

After the NRC panel finished their remarks, the meeting moved to comments and questions from the public. I was the third member of the audience to speak to the NRC. Figure 3 shows Brett Klukan at the podium to the left, the NRC panel in the center, and several members of the audience turning to look at the speaker standing at the microphone located towards the back of the room out of view to the far right.

Fig. 3 (Source: Nuclear Regulatory Commission)

I asked the NRC four questions. After I posed the four questions, the NRC panel answered. My questions and the NRC’s answers:

UCS Question #1

The NRC’s 20-member inspection team covered a lot of ground, but still examined a small fraction of the safety systems at Pilgrim. Based on the large number of safety violations in the small sample the team examined, what assurance can the NRC provide about the state of the majority of safety systems the team did not examine?

NRC Answer: The NRC’s reactor oversight process (ROP) features periodic inspections of safety systems at Pilgrim with the team inspection being supplemental to those activities. If there were problems in those other safety systems, the periodic inspections would reveal them.

UCS Response: Don Jackson described his team identifying 10 to 15 apparent violations of federal safety regulations in the small sample of safety systems they examined—violations that apparently were NOT revealed previously by the ROP’s periodic inspection efforts. Those routine inspection efforts failed to identify violations among the small sample, strongly suggesting that the routine inspection efforts also fail to find violations in the larger sample.

UCS Question #2

Don Jackson explained that the text in his email about the staff at Pilgrim appearing overwhelmed or shocked referred to their reaction to the arrival of the NRC’s 20-member inspection team. Does the NRC believe that this staff might also be overwhelmed or shocked in response to an accident?

NRC Answer: Don Jackson explained that his email comments referred primarily to the plant’s support staff (e.g, engineers, maintenance workers, etc.) rather than about the control room operators. Don said that his assessment of the operators at Pilgrim during their duties in the control room and during exercises on the control room simulator gave him complete confidence that the operators would be able to successfully respond to an accident.

UCS Response: Even if Don’s assessment is correct (and the operators losing control of the reactor during a routine startup causing it to automatically shut down to avoid fuel damage, the operators mis-operating numerous safety components following Winter Storm Juno and the operators not receiving proper training on the use of the high pressure coolant injection system leaves room for doubt), it is incomplete. The response to an accident involves considerably more than the handful of operators on duty at the time. NRC’s regulations require dozens of other plant workers to staff the Technical Support Center, the Operations Support Center, and the Emergency Operations Facility. The work force freaking out because 20 NRC inspectors arrive on site—by an appointment made weeks in advance—suggests that work force could be equally stressed out responding to an unannounced accident.

UCS Question #3

Dan Dorman mentioned the NRC planned to conduct another public meeting in late March about this inspection and to release the team’s final report in mid-April. Would it be possible for the NRC to issue the final report before the public meeting to allow the public to review the report and participate meaningfully in the meeting?

NRC Answer: Don Jackson mentioned that the report for a recent team inspection at another nuclear plant was over 350 pages due to all the information it contained. He said it would take sustained effort for the report by the team for their inspection at Pilgrim to be issued by mid-April, with no real opportunity for putting it out sooner.

UCS Response: There are two items both under full control of the NRC—the public meeting and the team inspection report. I have no reason to doubt Don’s word that mid-April is the soonest that the report can be released. I have every reason to doubt why the NRC must hold the public meeting in late March. The NRC could conduct the public meeting in late April, or early May, or mid-May, or late-May, or early June, or any time after they release the team’s report. The only reason for the NRC to conduct a public meeting about a non-existent report is because that’s the way they prefer to do it.

UCS Question #4

Audience members for this meeting are given three strikes before they are out of the meeting. How many strikes has the NRC given Pilgrim before it is out?

NRC Answer: Bill Dean began to answer the question, but Dan Dorman interrupted him. Dan labeled the question rhetorical and directed Brett to proceed with the next speaker.

UCS Response: I appreciate NRC bringing back Bert the turtle with this Duck and Cover gimmick. To be sure, I’d have better appreciated the NRC’s explanation why audience members get dragged out of the room after three strikes while Pilgrim does not get shut down after 10 to 15 violations of federal safety regulations. But this is America where everyone has the right to chicken out. My apologies if I put the NRC in a fowl mood.

To Be (Shut Down) or Not to Be (Shut Down)

The recurring theme during the meeting was whether the known performance problems warranted the shutdown of Pilgrim (either permanently or until the problem backlog was eliminated) or if Pilgrim could continue operating without exposing the community to undue risk.

Best I could tell, the meeting did not change any participant’s viewpoint. If one entered the room believing Pilgrim was troubled but sufficiently safe, one left the room with this belief intact. If one entered the room feeling Pilgrim’s problems posed too great a hazard, one probably left the room with even stronger convictions.

The meeting was somewhat like a court trial in that two reasonably supported but entirely opposite arguments were presented. The meeting was unlike a court trial in that instead of a jury, only time may decide which argument is right.

The Argument for Pilgrim Continuing to Operate

The team inspection led by Don Jackson is a direct result of an increasing number of problems at Pilgrim that caused the NRC to drop its performance assessment from Column 1 of the ROP’s Action Matrix into Column 2, 3 and eventually 4. The NRC developed the ROP in the late 1990s in response to high-profile troubled nuclear plants like Millstone, Salem, and Cooper.

The Action Matrix has five columns. A reactor with performance so bad that the NRC places it into Action Matrix Column 5 cannot operate until the NRC is satisfied enough of the problems have been corrected to permit restart.

Dan Dorman and Don Jackson tried to explain during the meeting that it was not the number of problems that determined placement into Column 5, it was the severity of the problems that mattered. They said several times that the 10 to 15 apparent violations identified by the team reinforced the NRC’s determination that Pilgrim was a Column 4 performer, but did not cause them to feel movement into Column 5 was warranted.

The Action Matrix is like our legal system. Persons guilty of a single misdemeanor generally receive lesser sanctions than persons guilty of multiple misdemeanors who in turn generally receive lesser sanctions than persons guilty of a single felony. Persons guilty of multiple felonies tend to be those receiving the severest sanctions and incarceration.

Pilgrim got into Column 4 as the result of several violations identified by NRC inspectors that were classified as White, the second least severe classification in the NRC’s Green, White, Yellow, and Red system. The data suggest performance shortcomings warranting regulatory attention, but it doesn’t suggest a trip to nuclear jail.

The Argument for Pilgrim Shutting Down

The NRC panelists stated several times during the meeting that they did not see any immediate safety concern that required Pilgrim to be shut down. Those assurances would be more meaningful and credible had the panelists or their NRC colleagues periodically seen an immediate safety concern, even from a distance.

The last time the NRC saw an immediate safety concern and ordered an operating reactor to shut down was March 31, 1987 when the agency ordered the Unit 2 and 3 reactors at the Peach Bottom nuclear plant in Pennsylvania to be shut down (the Unit 1 reactor had already been permanently shut down). Dan Dorman and Ray Lorson did not join the NRC staff until 1991. Don Jackson did not come to the NRC until 2003. Of the four NRC panelists, only Bill Dean was with the agency the last time an immediate safety concern was spotted.

Yet there have been times since 1987 when immediate safety concerns have existed:

Davis-Besse Safety Blindspot

In the fall of 2001, the NRC staff drafted an order that would require the Davis-Besse nuclear plant to be shut down. To justify the order, the NRC staff assembled the strongest circumstantial case one could hope to build that an operating reactor was unsafe. The NRC staff evaluated the reactor against five criteria in Regulatory Guide 1.174 (RG 1.174). All five criteria had to be satisfied for a reactor to be considered safe. The NRC staff determined that one criterion was not met and the other four criteria were most likely not met. Absent dead bodies or a mushroom cloud, you cannot build a stronger case that an operating reactor is unsafe.

Fig. 4 (Source: Nuclear Regulatory Commission)

But NRC senior managers shelved the order and allowed Davis-Besse to continue operating. When the reactor finally shut down, workers discovered the reactor was less safe than the NRC staff had feared. Per the NRC, Davis-Besse came closer to a meltdown than any reactor since the Three Mile Island accident in March 1979 (much closer than Peach Bottom ventured in March 1987).

Worse still, when interviewed by the NRC’s Office of the Inspector General, the NRC senior managers stated, under oath, stood behind their decision. They claimed they needed absolute proof that an operating reactor was unsafe before they would order it shut down. Somehow, failing to meet five of five safety principles does not constitute absolute proof to the NRC. Perhaps not meeting eight or nine out of five safety principles would suffice.

Oconee Safety Blindspot

In June 2010, the NRC issued a confirmatory action letter (CAL) to the owner of the Oconee nuclear plant in South Carolina. The CAL required that the owner take fifteen steps to reduce risk of failure at the upriver Jocassee Dam (which was also owned by Oconee’s owner) and to lessen the flooding vulnerability at Oconee should the dam fail.

The NRC staff discovered that the failure rate for the Jocassee Dam was as high as other hazards that Oconee was protected against. Thus, failure of the dam could not be dismissed as incredible or overly speculative.

The NRC staff further estimated that if the Jocassee Dam failed, flooding at the Oconee site created a 100 percent chance of causing all three operating reactors to melt down, all cooling of the spent fuel pools to be lost, and all three reactor containments to fail.

The high risk of flooding causing three operating reactors to melt down prompted the NRC to issue the CAL to Oconee’s owner nine months before flooding caused three operating reactors at Fukushima to melt down.

The hazard was real enough to cause NRC to require the owner to take steps to lower the risk, but not real enough to warrant the reactors to shut down until the risk was better managed.

Most galling is the fact that the NRC withheld information about this hazard from the public. Their June 2010 CAL was issued in secret. When the NRC conducted their annual public meeting in the Oconee community in April 2011—about six weeks after flooding melted three operating reactors at Fukushima—they said nothing about the CAL being issued to better manage flooding vulnerabilities at Oconee. The public cannot trust an agency that withholds relevant information from them.

It may be true that the NRC would order an operating reactor to be shut down if it saw an immediate safety concern. But it’s been nearly thirty years since the NRC noticed an immediate safety concern at an operating reactor. Since then, the NRC has noticed very serious safety problems at Davis-Besse and Oconee, yet allowed those reactors to continue operating.

The Davis-Besse and Oconee cases occurred after the NRC adopted the ROP and its Action Matrix. None of the safety problems that led to the NRC staff drafting a shutdown order for Davis-Besse or issuing a CAL for flood protection problems at Oconee were considered in the ROP. Thus these safety problems were entirely invisible as far as the Action Matrix was concerned.

The NRC should not rely on a safety yardstick that ignores significant safety issues.

UCS’s Argument about Pilgrim

Because the NRC has demonstrated its ability to jettison safety standards when an operating reactor doesn’t measure up, and because it has not recently demonstrated an ability to spot an immediate safety concern, it is entirely reasonable for the community around Pilgrim to have anxiety about the plant’s known performance problems. Shutting down Pilgrim would lessen that anxiety.

Should public anxiety be used as a pretext for shutting down an operating reactor?

Absolutely not.

Instead, the public should have trust and confidence in the NRC to protect them from Pilgrim’s problems. But the NRC has not done much to warrant such trust and confidence by the NRC. If public anxiety is high, it’s because public trust and confidence in the NRC is low.

Public trust and confidence in the NRC should be the proper context for a troubled reactor continuing to operate.

That proper context is missing.

The NRC must take steps to restore public trust and confidence. They should consistently establish and enforce safety regulations. NRC senior managers must stop looking for absolute proof that operating reactors are unsafe and instead look for absolute proof that operating reactors comply with federal safety regulations.

And when NRC senior managers see safety problems, they must disclose that finding to the public. Hiding such information, as they did with the flooding vulnerabilities at Oconee, provides the public with a distorted view. And such antics provide the public with zero reason to trust anything the NRC utters. When you cherry-pick what you say and when you say it, you stop being a credible authority.

If the NRC allows Pilgrim to continue operating and the reactor has an accident, will the agency be able to honestly look victims and survivors in the eye and say they did everything they could to protect them?