Following the March 1979 reactor core meltdown at the Three Mile Island nuclear plant in Pennsylvania, the US Nuclear Regulatory Commission (NRC) established a safety policy that sought to limit the chance of another meltdown to no more than once every 10,000 years of reactor operation—reasonably remote odds for a reactor licensed to operate for 40 years. But since that safety goal was established, the NRC has extended the operating licenses of more than three quarters of the US fleet of 100 reactors by 20 years and is contemplating extending the licenses for an additional 20 years. The new license process, called Subsequent License Renewal, would extend operations from 60 years to 80 years. Although some reactors in unregulated markets have retired early because they can’t compete economically with cheap natural gas, reactors in regulated markets face a very different set of economic circumstances and may be kept in service well past their originally planned retirement dates.
The chance of one reactor experiencing a meltdown among a fleet of 100 reactors operating within the NRC’s safety goal for 40 years is nearly one in three (32.97 percent), or slightly higher than the risk from taking two turns on a six-chamber revolver during Russian roulette. The chance of a meltdown from that fleet operating for 60 years rises to 45.12 percent, or slightly higher than taking three Russian roulette turns. And the meltdown risk from the fleet operating for 80 years is 55.07 percent, or roughly the risk from taking four and one-half Russian roulette turns.
The bathtub curve. The NRC’s safety goal is a constant number for all reactors at every point during their operation. In reality, the risk over a reactor’s lifetime varies by what is called the bathtub curve due to its shape.
A reactor begins operating with relatively high risk due to material imperfections, assembly errors, worker mistakes, and other break-in problems. The risk levels off during mid-life and then rises late in life due to age-related degradation.
The US fleet of reactors is heading toward, if not already in, the wear-out portion of the bathtub curve where risk increases. In addition, the five new reactors (Watts Bar Unit 2, Vogtle Units 3 and 4, and Summer Units 2 and 3) about to join the fleet cannot skip forward to the middle-age period of relatively low risk—they must navigate through the high-risk break-in phase.
Power companies deciding whether to extend the lifetimes of their existing nuclear reactors or to replace them with new reactors confront the reality of the bathtub curve. Recent examples illustrate that well-intended decisions can go awry:
Crystal River 3. In December 2008, the owner of Crystal River 3 in Florida applied to the NRC for a 20-year extension to the reactor’s operating license. In fall 2009, the reactor shut down for a planned refueling outage. The tasks scheduled during this outage included replacing the steam generators. The original components were wearing out, evidenced by plugged tubes. The replacement steam generators would restore safety margins, supporting the reactor’s operation throughout the renewed license period.
That was the intention. The reality was that cutting through the concrete containment structure’s wall to get the old steam generators out, and to install the new ones, damaged the concrete. Initial efforts to repair the damage failed, and the cost of fixing the damage was too high to justify. The owner opted to permanently shut down the reactor, and withdrew its license renewal application.
San Onofre. The owners of the San Onofre Units 2 and 3 reactors in southern California also elected to replace the original steam generators to restore safety margins during the remainder of their 40-year licenses and to support license renewal should the company decide to pursue that option. The steam generators were successfully replaced, and both reactors returned to service, but not for long. On January 31, 2012, workers shut down the Unit 3 reactor after one of the tubes inside a replacement steam generator broke. The Unit 2 reactor was already shut down so that workers could inspect for, and plug, damaged tubes inside its replacement steam generators.
The replacement steam generators had design and manufacturing problems that caused their tubes to wear out much more rapidly than expected. In other words, the replacement steam generators experienced break-in failures—lots of them. The owner opted to permanently shut down the reactors rather than incur the cost of fixing the flawed steam generators or replacing the replacements.
Fort Calhoun. The Fort Calhoun reactor in Nebraska began operating in 1973. The NRC renewed its 40-year operating license for an additional 20 years in November 2003 after determining that aging-management programs were in place to sustain safety margins over the longer period.
Consistent with those programs, workers replaced parts of valves during the spring 2015 refueling outage. The rubber parts were vulnerable to radiation-induced damage, so workers installed replacement parts less vulnerable to radiation exposure. Within hours of the reactor’s startup after the refueling outage, however, the refurbished valves failed to move. The replacement parts had degraded as a result of the high temperatures they encountered during reactor operation, essentially gluing the valves in place.
The bathtub curve had claimed another victim. Workers had replaced parts within the valves because the original parts were susceptible to an aging mechanism that hastened their entry into the wear-out zone. Their efforts to avoid component failure were defeated when the replacement parts were even more susceptible to another degradation problem, preventing the rebuilt valves from getting out of the break-in phase.
Browns Ferry Unit 1. Recent decisions about whether to continue operating 1970s-vintage reactors rather than replacing them with 21st century designs strongly suggest that claims made in slick marketing brochures about safe, economical, reliable reactors are being viewed with healthy skepticism in corporate boardrooms. The Browns Ferry Unit 1 is a good example of an aging reactor winning out over new construction.
The Tennessee Valley Authority (TVA) shut down all three reactors at the Browns Ferry Nuclear Plant in March 1985 due to numerous safety issues. TVA restarted the Unit 2 and 3 reactors after years of repairs, but decided not to restart Unit 1. Two decades later, faced with restarting Unit 1 or building new reactors at its partially constructed Bellefonte nuclear plant, which is also in northern Alabama, the TVA Board authorized nearly $2 billion to restart the 43-year-old reactor. Unit 1 restarted in 2007. TVA recently cancelled plans to construct and operate new reactors at Bellefonte, and might sell the site.
Watts Bar Unit 2. TVA began constructing two reactors at the Watts Bar plant in Tennessee in the 1970s. Delays prevented the Unit 1 reactor from going into service until 1996. TVA halted construction on Unit 2 for many years but finally authorized more than $4 billion to resume and complete construction. Unit 2 is slated to begin operating in mid-2016.
The economic counterpart to the safety bathtub curve likely factored into TVA’s decision to invest in old reactors rather than in new ones. Browns Ferry Unit 1 and Watts Bar Unit 2 are virtually identical to other reactors operated by TVA. Therefore, TVA’s leaders could place greater certainty in what they would get for their investment in old technology than what they might get from an equal, or larger, investment in 21st century reactors.
None of the five reactors currently under construction in the United States has yet split its first atom. Thus, no evidence exists to indicate whether the new reactor designs have successfully incorporated lessons learned from the past so as to reduce the height and width of the break-in portion of the bathtub curve, or whether surprises yet to be revealed will keep the break-in curve at the existing level or even higher. The significant schedule delays and large cost over-runs for reactors under construction in Georgia, South Carolina, Finland, and France illustrate the economic risks of building new, untried reactors.
Nuclear power in the future. On paper, nuclear reactors pose no risks. But when they move from blueprints to backyards, they pose very real safety and economic risks that must be managed; otherwise, safety levels drop and costs rise unnecessarily.
Many new and existing US reactors are operating in the high-risk break-in and wear-out portions of the bathtub curve. The nuclear industry and its regulator must be aggressively vigilant to keep the factors responsible for these higher risks in check as much as possible.
In theory, refurbishing or replacing parts can safely extend the lives of nuclear reactors well beyond their originally envisioned life spans—perhaps even doubling longevity to 80 years. But in practice, refurbishment and replacement may merely swap wear-out failures for break-in failures. The nuclear industry and its regulator must devote more resources to this issue, especially as aging US nuclear reactors require more and more upkeep. To balance the risks of operating reactors for 60 years or longer, the NRC’s license renewal process must include substantive risk-reduction measures.
Editor’s note: David Lochbaum prepared this article on his own time, and it may not represent the views of the Union of Concerned Scientists.