Fixing the Nuclear Regulatory Commission

The United States operates 104 nuclear power reactors, which provide nearly 20 percent of the nation’s electricity. More than half have had their original 40-year operating licenses renewed for an additional 20 years. Encouraged by billions of dollars in subsidies and incentives in the 2005 Energy Bill, a handful of companies applied for licenses to build new reactors last fall, and other companies are expected to apply later this year. Recurring lessons from the past consistently inform us that unless the Nuclear Regulatory Commission (NRC) undergoes major reforms, nuclear power will remain both riskier and more expensive than necessary.

The NRC is the federal agency primarily responsible for establishing and enforcing safety regulations for nuclear power. It does the former well. It does the latter poorly. The Union of Concerned Scientists (UCS) has monitored nuclear power safety issues since the early 1970s. We have seldom argued that the NRC needed to raise its safety standards. Instead, we have almost always contended the safety bar provided appropriate management of risk, but that one or more nuclear plants was doing the limbo beneath it. Most of our efforts have been directed at getting the NRC to enforce regulations already on the books.

Evaluations conducted by the Government Accountability Office (GAO) and the NRC’s Inspector General (IG) confirm our perspective: These reports repeatedly identify inadequate enforcement of existing regulations by the NRC. For example, in its May 2004 report, “Nuclear Regulation: NRC Needs to More Aggressively and Comprehensively Resolve Issues Related to the Davis-Besse Nuclear Power Plant’s Shutdown”, the GAO concluded, “[The] NRC should have but did not identify or prevent the corrosion at Davis-Besse [a nuclear power plant in Ohio] because both its inspections at the plant and its assessments of the operator’s performance yielded inaccurate and incomplete information on plant safety conditions.”

The IG’s January 2008 report, “NRC’s Oversight of Hemyc Fire Barriers”, documents the NRC’s repeated failure to enforce fire-protection regulations. In March 1993, after problems surfaced with the Thermo-Lag fire barrier used by nearly 100 reactors, the NRC chairman committed to evaluate all fire barriers used in U.S. nuclear reactors. Tests conducted by the National Institute of Standards and Technology in 1993 (and reported to the NRC in 1994) found that the one-hour Hemyc fire barrier, used by 17 nuclear reactors, failed in 23 minutes. The NRC considered these tests too small to be conclusive and stated that larger-scale testing was needed. However, it wasn’t until 2005 that the NRC commissioned such testing–even though the NRC acquired yet more evidence of problems with Hemyc in 2000. After an inspection found that Hemyc was used more extensively than assumed at one U.S. plant, the NRC reviewed the Hemyc tests conducted by the vendor and found that they did not demonstrate that Hemyc could meet its one-hour or three-hour ratings. When the larger-scale tests were finally conducted by Sandia National Laboratory, the one-hour Hemyc fire barrier failed in 13 minutes.

According to the IG: “As of December 2007, no fire-endurance tests have been conducted to qualify Hemyc as an NRC-approved 1-hour or 3-hour fire barrier for installation at [nuclear power plants].” Thus, the NRC has known since 1994 that 17 U.S. reactors are relying on Hemyc for fire protection and that Hemyc does not meet NRC standards, but has not enforced the regulations it established in 1980 as a result of the serious fire at the Browns Ferry nuclear plant in Alabama that disabled the power, control, and instrumentation cabling for all the emergency core cooling systems on Unit 1 and most of those systems on Unit 2. The regulations included requirements that cabling for primary and backup safety systems (a) be physically separated by at least 20 feet horizontally, or (b) be protected by a one-hour or three-hour fire barrier to lessen the risk that a single fire disables all emergency systems.

One could contend that the GAO and IG investigators are predisposed to finding inadequate enforcement rather than challenging the adequacy of the underlying regulations. After all, auditors tend to be process-oriented; their task is to assess how well the organization is doing what it claims to be doing. However, the NRC’s own assessments of its regulatory meltdowns also repeatedly conclude that the majority of problems stem from inadequate enforcement of adequate regulations.

For example, the NRC lessons-learned task force examined the regulatory failures associated with the near-accident at Davis-Besse in 2002 ^* and made 49 recommendations for actions the NRC should take to prevent recurrences. Forty-six of these outlined ways to improve enforcement of existing regulations, while the remaining three dealt with upgrading the underlying regulations. The NRC’s lessons-learned efforts for Indian Point (New York), Millstone (Connecticut), South Texas Project, and other troubled nuclear plants provide similar findings–the regulations are not the problem, enforcement is.

NRC’s inadequate enforcement has caused significant safety and economic problems. In its September 2006 report, “Walking a Nuclear Tightrope: Unlearned Lessons of Year-plus Reactor Outages,” UCS described the 36 times since 1966 that U.S. nuclear power reactors remained shut down a year or longer to restore safety levels eroded by accumulated violations. In these cases, it took an army of workers more than a year, and cost an average of nearly $1.7 billion, to bring the reactor back into compliance. Inadequate enforcement by the NRC allowed safety levels to erode over several years, resulting in unnecessarily higher risk to the surrounding communities during those years and higher cost to the owners.

If UCS, GAO, IG, and NRC all identify inadequate enforcement of safety regulations as the root cause of NRC’s regulatory breakdowns, why hasn’t the problem been fixed? Quite simply, while the NRC suffers from the same affliction that impaired performance at the reactors involved in the regulatory breakdowns, the NRC has never received the same cure applied at those sites. Consequently, the NRC remains mired in a performance rut.

The 10 nuclear reactors at Davis-Besse, Cook (Michigan), LaSalle (Illinois), Clinton (Illinois), Millstone, and Salem (New Jersey) have experienced year-plus outages in the past decade. At each site, new senior managers were brought in to be agents of change by breaking bad behavior patterns and instilling the proper safety culture. NRC inspectors swarmed the reactors to verify that the reforms promised by the new managers were achieving the advertised gains.

UCS, GAO, IG, and NRC issued reports on the associated regulatory breakdowns at Salem, Millstone, and Davis-Besse. But unlike the sites, new managers were not brought in to NRC to be agents of change. Instead, the NRC’s agents of status quo “resolved” the reports’ recommendations with changes that failed to remedy the root causes of NRC’s performance problems. And congressional investigators did not swarm around NRC’s offices to verify that the improvements recommended by GAO and IG had been implemented at all, let alone adequately.

The Salem, Millstone, and Davis-Besse reactors are operating at substantially higher performance levels today than prior to their extensive, monitored reforms while the NRC operates at essentially the same performance level today as it has for decades. When an NRC senior manager retires or resigns, everyone moves up one rung on the ladder. The “new” senior manager has typically been at NRC for two decades. There may be a new face behind the desk, but there’s not new blood, new perspectives, or new approaches. If the NRC is to make tangible improvements in how it enforces its regulations, it needs new managers who can be agents of change. These individuals can think outside the box because they did not spend the past two decades confined inside the box.

UCS is not advocating a purge of senior managers at NRC followed by busloads of replacements brought in from “Managers ‘R Us” or eBay. Instead, UCS thinks a two-pronged approach would work just fine. First, as normal attrition creates openings at the senior management level, the NRC must fill these openings with the best available candidates, not just the best candidate from within the NRC. To enhance the skill set of NRC’s middle managers, the NRC–with the help of Congress–should develop a rotational program with sister agencies. Middle managers at NRC could be loaned to or exchanged with those at, for example, the Energy Department, EPA, and GAO to bring back outside perspectives.

Congress needs to play an active, sustained role in NRC reforms. Just as the NRC actively monitored the reform activities at Davis-Besse and Millstone, Congress must watch over the NRC’s efforts to ensure that the agency attains its reform goals in a timely manner.

With these reforms in the rearview mirror instead of on the road ahead, the NRC will become as good at enforcing its regulations as it was in creating them. As a result, safety levels at nuclear power plants will be where they are supposed to be, stopping the seemingly endless string of costly outages needed to restore eroded safety levels.

Of course, NRC need not undergo the reforms now. The reforms can be deferred until after the next nuclear plant disaster using the precedent applied at NASA after Columbia, the intelligence community after 9/11, and FEMA after Katrina. The reforms will be the same; their cost will be significantly higher.

^*According to the NRC, Davis-Besse came closer to an accident than any reactor since Three Mile Island. A crack formed in a metal tube entering the reactor vessel’s lid and leaked borated water onto the carbon steel. The boric acid residue ate completely through the 6-inch carbon steel vessel to expose a one-quarter-inch stainless steel cladding applied to the vessel’s inner surface. The timeline spanned an estimated six years and provided numerous opportunities for the NRC to step in. In the last missed opportunity, NRC staff drafted an order requiring Davis-Besse to shut down immediately on the basis that the reactor failed to satisfy four of the agency’s five safety criteria and probably did not meet the fifth. But NRC’s senior managers shelved the draft order because it would have cost the company too much money and instead waited to inspect the reactor for several months until it had a scheduled shutdown for refueling.