Reactor Design in Japan Has Long Been Questioned

By Tom Zeller Jr., The New York Times

15 March 11

he warnings were stark and issued repeatedly as far back as 1972: If the cooling systems ever failed at a Mark 1 nuclear reactor, the primary containment vessel surrounding the reactor would probably burst as the fuel rods inside overheated. Dangerous radiation would spew into the environment.

Now, with one Mark 1 containment vessel damaged at the embattled Fukushima Daiichi nuclear plant in Japan and other vessels there under severe strain, the weaknesses of the design - developed in the 1960s by General Electric - could be contributing to the unfolding catastrophe.

When the ability to cool a reactor is compromised, the containment vessel is the last line of defense. Typically made of steel and concrete, it is designed to prevent - for a time - melting fuel rods from spewing radiation into the environment if cooling efforts completely fail.

In some reactors, known as pressurized water reactors, the system is sealed inside a thick steel and cement tomb. Most nuclear reactors around the world are of this type.

But the type of containment vessel and pressure suppression system used in the failing reactors at Japan's Fukushima Daiichi plant - and in 23 reactors at 16 American plants - is physically less robust, and has long been thought to be more susceptible to failure in an emergency than competing designs.

G.E. began making the Mark 1 boiling-water reactors in the 1960s, marketing them as cheaper and easier to build - in part because they used a comparatively smaller and less expensive containment structure.

American regulators began identifying weaknesses very early on.

In 1972, Stephen H. Hanauer, then a safety official with the Atomic Energy Commission, recommended in a memo that the sort of "pressure suppression" system used in G.E.'s Mark 1 plants presented unacceptable safety risks and that it should be discontinued. Among his concerns were that the smaller containment design was more susceptible to explosion and rupture from a buildup in hydrogen - a situation that may have unfolded at the Fukushima Daiichi plant.

"What are the safety advantages of pressure suppression, apart from the cost saving?" Mr. Hanauer asked in the 1972 memo. (The regulatory functions of the Atomic Energy Commission were later transferred to the Nuclear Regulatory Commission.)

A written response came later that same year from Joseph Hendrie, who would later become chairman of the N.R.C. He called the idea of a ban on such systems "attractive" because alternative containment systems have the "notable advantage of brute simplicity in dealing with a primary blowdown."

But he added that the technology had been so widely accepted by the industry and regulatory officials that "reversal of this hallowed policy, particularly at this time, could well be the end of nuclear power."

In an e-mail on Tuesday, David Lochbaum, director of the nuclear safety program at the Union of Concerned Scientists, said those words seemed ironic now, given the potential global ripples on the nuclear industry from the Japanese accident.

"Not banning them might be the end of nuclear power," said Mr. Lochbaum, a nuclear engineer who spent 17 years working in nuclear facilities, including three that used the G.E. design.

Questions about the G.E. reactor design escalated in the mid-1980s, when Harold Denton, an official with the N.R.C., asserted that Mark 1 reactors had a 90 percent probability of bursting should the fuel rods overheat and melt in an accident. A follow-up report from a study group convened by the commission concluded that "Mark 1 failure within the first few hours following core melt would appear rather likely."

In an extreme accident, that analysis held, the containment could fail in as little as 40 minutes.

Industry officials disputed that assessment, saying the chance of failure was only about 10 percent.

Michael Tetuan, a spokesman for G.E.'s water and power division, staunchly defended the technology this week, calling it "the industry's workhorse, with a proven track record of safety and reliability for more than 40 years."

Mr. Tetuan said there were currently 32 Mark 1 boiling-water reactors operating safely around the globe. "There has never been a breach of a Mark 1 containment system," he said.

Several utilities and plant operators also threatened to sue G.E. in the late 1980s after the disclosure of internal company documents dating back to 1975 suggesting that the containment vessel designs were either insufficiently tested or had flaws that could compromise safety.

Paul Gunter, director of the Reactor Oversight Project with Beyond Nuclear, an organization opposed to nuclear power, said regulators and utilities began raising concerns about the containment design as far back as the 1970s.

"The key concern has always been that the containment structure was undersized, and that a potential accident could overwhelm and rupture it," Mr. Gunter said.

The Mark 1 reactors in the United States have undergone a variety of modifications since these initial concerns were raised. Among these, according to Mr. Lochbaum, were changes to the doughnut-shaped torus - a water-filled vessel encircling the primary containment vessel that is used to reduce pressure in the reactor. In early iterations, steam rushing from the primary vessel into the torus under high pressure could cause the vessel to literally jump off the floor.

In the late 1980s, all Mark 1 reactors in the United States were also ordered to be retrofitted with venting systems to help reduce pressure in an overheating situation, rather than allow it to build up in a containment system that regulators were concerned could not take it.

It is not clear precisely what modifications were made to the Japanese boiling-water reactors now failing, but James Klapproth, the chief nuclear engineer for General Electric Hitachi, said a venting system was in place at the Fukushima plants to help relieve pressure.

With electrical power cut off in the aftermath of the earthquake and backup sources of power either failing or exhausted, workers have been struggling to inject seawater into the reactor to maintain control, but they have had some trouble venting the resulting steam.

Mr. Gunter argued that in any event, such venting amounts to a circumvention of the whole notion of containment in the first place. "They essentially have to defeat containment to save it," he said.

What role the specifics of the G.E. design is playing in the rapid deterioration of control at the Fukushima plant is likely to be a matter of debate, and it is possible that any reactor design could succumb to the one-two punch of an earthquake and tsunami like those that unfolded last week in Japan.

Although G.E.'s liability would seem limited in Japan - largely because the regulatory system in that country places most liability on the plant operator - the company's share price was down more than 2 percent at midday Tuesday as the situation at the Fukushima plant deteriorated.

Still, Mr. Lochbaum said it was important to emphasize that the design specifications for containment and cooling on any reactor are a matter of balance. The primary alternate reactor design, the pressurized water reactor, calls for a thicker and bigger containment structure, for example. A boiling-water reactor design like the one at Fukushima does allow for scaling back on the size of the containment system while ostensibly maintaining the requisite safety margins.

In that sense, Mr. Lochbaum said, G.E.'s boiling-water reactors should be no better or worse in weathering accidents than any other design.

Should the ability to cool the reactor completely fail, however, Mr. Lochbaum said, "I'd certainly rather have a bigger, thicker containment structure."