The NRC at Three Mile Island

First of a three post series on the Three Mile Island meltdown

Figure 1. Portion of Conrad Cartoon in LA Times day after NRC says hydrogen bubble could explode.

Most of the descriptions of Three Mile Island focus on the technical problems that resulted in the meltdown: the design faults, the operator mistakes, the blinkered training that instructed them to make those mistakes, despite the fact that nearly the same scenario 18 months earlier had revealed the training was both wrong and dangerously misleading. Excellent summaries can be found in Ted Rockwell's book, Creating the New World, and at Mike Derivan's web site. This piece focuses on the NRC's role in turning a very expensive industrial casualty that measurably harmed no one, into a clarion call to shut down nuclear power.

Prologue

Table 1 compares the amount of the most harmful isotopes released at the four largest nuclear reactor releases to date.

Table 1. Four Main Isotopes at Four Big Releases.

The size of the release at Three Mile Island was over 1000 times smaller than the release at Windscale. The GKN argues that to first order the harm associated with a release goes roughly as the square of the size of the release. By this harm metric, TMI was at least a million times less harmful than Windscale. But at Windscale, there was no panic, no evacuation, and no measurable harm to the public. The loss was a crappy nuclear weapons reactor and some milk.

At TMI, there was widespread panic. Bishop Keeler of Harrisburg was so convinced his flock faced imminent annihilation that he declared general absolution. Over 100,000 people evacuated. The media went wild. Images of an atom bomb explosion spread across the nation. There was widespread anguish, anger, and a 20 percent swing in public support of nuclear power. What was the difference? That's an easy one. It was the Nuclear Regulatory Commission (NRC).

Wednesday, March 28

At 4:00 AM on March 28th, the TMI2 turbine shut down unexpectedly. Due to a series of mechanical failures, and wrong responses by operators misreading the situation as a result of poor control room design and flat wrong training, this easily handled upset escalated into a casualty that resulted in the destruction of the reactor core and a multi-billion dollar write off. But by 6:22 AM, the crew had discovered and corrected the key problem. However, the loss of cooling water was such that a portion of the core had been uncovered; decay heat drove fuel element temperatures to 2200C, and a large portion of the core melted down.¹ By 7:50 PM that evening cooling was restored, and the situation was back under control. That should have been the end of it.

But radioactive gas had leaked from the melted fuel elements to the cooling water, and then outside containment via a stuck open valve. Some of that gas escaped to the environment. Well above background dose rates were measured outside the plant. The NRC decided not to admit there had been a release; and lied in the following statement put out in time for the 7 PM news.

Low levels of radiation have been measured off the plant site. The maximum confirmed radiation reading was about three milliroentgen per hour (0.03 mGy/h) about one-third mile from the site. ... It is believed that this is principly direct radiation coming from radioactive material with the reactor compartment, rather than from release of radioactive materials from the containment.

The NRC concocted that whopper to protect the Negligible Probability Lie. Hard to imagine a stupider move. Here's how NBC's Tom Brokaw interpreted this:

The Nuclear Regulatory Commission in Washington says radiation penetrated through walls that were four feet thick, and it spread as far as 10 to 16 miles from the plant.

Brokaw may have embellished a bit, but he had an impeccable source. At least 3 of the 5 NRC commissioners cleared this nonsensical falsehood.\cite{rogovin-1980}

TMI was now national news. The New York News ran a headline ``Nuke Plant Spews Radiation in Pa., Goes thru 4 ft walls".\cite{walker-2004}[p 98] 400 media people flocked to the plant.

Thursday, March 29

Thursday the 29th was relatively calm, with the core numbers slowly improving. But there were two other problems.

1) The radioactive gas that had leaked and was leaking into portions of the plant would have to be dealt with.

2) During the period when the core had not been covered, the 2000C plus temperature resulted in the zirconium fuel element cladding reacting with the oxygen in the water, freeing up hydrogen which formed a bubble in the top of the reactor pressure vessel. Enough hydrogen in the primary loop could interfere with the cooling.² The bubble would need to be removed.

Friday, March 30

Early on Friday the 30th, gas pressure had built up in the makeup tank to the point that venting to the outside was necessary. The NRC in DC did a quick calculation of the ground level, site boundary dose rate from venting, and came up with 12 mGy/h. But they used a venting scenario very different from the actual. For the actual, the ground level number is at least 100 times smaller. At 7:10 AM venting started. At the same time, a helicopter just above the vent stack, measured a peak rate of 12 mGy/h. This was communicated to NRC DC, who misinterpreted it as confirmation for their estimate.

On the basis of a bad number and misread measurement, and without any attempt to confirm either, NRC-DC decided to call Pennsylvania Governor Thornburgh and recommend evacuation up to 10 miles downwind. Harold Denton, the NRC employee who made the decision later said: ``my sole objective was to minimize the radiation exposure to the public. I did not give any weight to whatever hardship evacuation might cause".\cite{walker-2004}[p 126] Denton told Harold Collins to call the state. Collins unlike Denton knew that the 12 mGy/h reading was just above the stack. But he made the call anyway because ``things were going to hell in a handbasket". At 9:15 AM on the 30th Collins called the Governor's office and recommended evacuation up to 10 miles downwind.

Thornburgh was getting an entirely different story from the plant and the NRC people at the plant, who knew the situation was under control and improving. They were incensed by a call they had not been told about. Thornburgh, understandably confused, called NRC Commission chairman, Joseph Hendrie, who by this time knew the 12 mGy/h number was bogus. Hendrie said no evacuation is necessary; but, off the cuff, added people should stay indoors, and maybe pregnant women and small children should evacuate. Thornburgh issued an advisory recommending pregnant women and preschool children move out of the area within 5 miles of the plant and ordered schools within that area to close. About 3500 souls evacuated as a result of this advice.

Mid-afternoon NRC-DC held a press conference. When one reporter ask the NRC rep if a meltdown was possible, he replied that it was not likely, but it was possible. The meltdown had happened two days earlier; but this was the first time the NRC acknowledged even the possibility. UPI quickly ran a story about the imminent danger of a meltdown.

Saturday, March 31

Despite all the nonsense coming out of NRC-DC, the area immediately around the plant remained calm. Through Saturday core temperatures continued to drop and radiation levels at the plant downwind boundary were in the 0.002 to 0.005 mGy/h range. These are background levels in parts of the planet. The MetEd operators were reducing the volume of the hydrogen bubble.

There was no way to vent the reactor vessel directly. So one approach used the make up and let down system to circulate water through the reactor. Some of the hydrogen dissolved into this water, and was extracted by the make up water filtration system. This path resulted in radioactive gases getting into the auxiliary building and eventually being vented to atmosphere.

The other approach was to cyclically over-cool the water in the pressurizer with the spray system. Hydrogen would evolve during the hot part of the cycle and be vented out the top of the pressurizer. This approach put H2 into the containment and the possibility of a flammable combination of hydrogen and oxygen in that space, which had to be avoided. MetEd had acquired recombiners, which react the hydrogen and oxygen non-explosively, and was setting up to install them. The first recombiner went into operation on Monday afternoon.

The reporters at TMI, who were talking to the plant people and the NRC employees on site, knew the situation was under control. Their focus shifted. They sought out distraught locals, who were very much in the minority, and filed ``color" stories about the anguish they were undergoing. Saturday evening, many went to the best restaurants in Harrisburg, Santana's Seafood House and Lombardo's. Reporters tend to flock together, swap stories, eat well and drink a lot at events like these. However, festivities were rudely interrupted by news of an AP editors' advisory filed at 8:23 PM,

Urgent. The NRC now says the gas bubble atop the nuclear reactor shows signs of becoming potentially explosive. Story upcoming.

The story cited an anonymous NRC source, who claimed the hydrogen bubble could explode within two days. Stan Benjamin, the author of the article, never revealed his NRC source; but it was almost certainly Dr. Roger Mattson, Director of Systems Safety, who had been told to look into the possibility of an explosion. Mattson had argued within the NRC that the bubble could explode in two days. Mattson was among the first to recognize that there had been extensive core damage. Starting Friday, he had lobbied hard for immediate evacuation of the whole area.

Sunday, April 1

The news quickly spread and panic ensued. Bishop Keeler declared general absolution. Tens of thousands threw their kids in the car and got the hell out of there. The journalists themselves were terrorized. Few if any knew any thing about nuclear power. When the hydrogen explosion story spread, they joined the exodus. Their subsequent stories reflected their own fright.\cite{bukro-2023}[p 397-398] Anything that scared the hell out of a seasoned war correspondent must be really bad. Many stories on Sunday were worded in a way that evoked visions of an atom bomb explosion. The LA times highlighted a cartoon, Figure 1, showing a cooling tower with a mushroom cloud coming out the top. The NRC claim produced a genius piece of propaganda.

The claim was nonsense. Casey Bukro, a Chicago Tribune reporter, who was there and experienced what he called ``the night of terror", put it this way many years later

The nation had been terrified by a bogus exploding bubble, a hoax, a fumbling miscalculation by one of the NRC's masters of technology, which even he could not fully comprehend.\cite{bukro-2023}[p 402]

Figure 2. TMI2. For present purposes, the key point is that, while there was oxygen in the containment building, there was nil oxygen in the Reactor Pressure Vessel, the green vertical oblong in the lower, middle of the drawing.

There was nil oxygen in the reactor pressure vessel (RPV), nor could there be. Oxygen was being generated in the RPV by radiolysis, the splitting of water into hydrogen and oxygen by decay heat radiation; but at a far lower rate than when the reactor is creating power. During normal operation, standard practice is to maintain 1 ppm dissolved hydrogen in the primary loop to immediately convert any new oxygen back to water.\cite{nsac-1980}[p HYD-2] This is basic chemistry known to any light water reactor operator. The dissolved hydrogen in the TMI2 primary loop was well above 1 ppm throughout.

Mattson never publicly admitted personal responsibility, but later he did say.

The amount of concern was entirely undeserved. There was never any danger of a hydrogen explosion to that bubble. It was a regrettable error. ... It originated in the staff.\cite{corey-1979}[p 55]

Mattson went on to a lucrative career as a nuclear power safety consultant.

The record is not completely clear here; but it seems Mattson asked his staff and experts the wrong questions. He asked what would be the rate of oxygen evolution in the core using worst case assumptions? He asked what would the damage be if all the hydrogen in the RPV exploded? Apparently he never asked how rapidly would the hydrogen soak up any oxygen that did evolve? This scenario is unforgivable.³ At least a dozen people were involved. Did none of these highly trained engineers and scientists not know to ask the obvious, right question? Or were they just giving the boss the answer he wanted?

This inexplicable blunder was quickly contradicted by the NRC on site. Governor Thornburgh issued a statement at 11 PM Saturday saying there was no immediate danger.⁴ Jimmy and Rosalynn Carter arrived the next morning, Sunday, and toured the plant. Then Carter gave a short speech asking everybody to calm down. Eventually they did, but the damage had been done.

Epilogue

The NRC worked hard to delay the return to normalcy. The always polite Kemeny report says:

By late Sunday afternoon, NRC --- which was responsible for the concern that the bubble might explode --- knew there was no danger of a blast and that the bubble appeared to be diminishing. It was good news, but good news unshared with the public. Throughout Sunday, the NRC made no announcement that it had erred in its calculations or that no threat of an explosion existed. Governor Thornburgh was not told of the NRC miscalculation either.\cite{kemeny-1979}[p 133]

On Tuesday, NRC's Denton finally says ``the bubble has been eliminated for all practical purposes." He is about two days late. Asked why, he gives credit to the NRC and luck. ``I think it was a little bit because of our actions and maybe a little bit of serendipity." A local NRC guy who was in the control room almost from the beginning was far more honest. ``The hydrogen bubble did not miraculously disappear. It was systematically and professionally eliminated by MetEd operators."\cite{rogovin-1980}[p 87] Thankfully, the NRC never assumed control of the actual shutdown.

During the first week of April, Thornburgh repeatedly asked the NRC if the pregnant women and preschool children could return home. But in the words of the President's Commission ``the NRC wanted some specific event as symbol that the crisis had ended".\cite{kemeny-1979}[p 138] The event they chose was cold shutdown which was weeks away. Cooler heads eventually prevailed. Governor Thornburgh lifted the advisory on April 7.

The NRC still refused to say the crisis was over, and cautioned that an evacuation might still be required. It was not until April 29th, that the NRC declared that the reactor was safely shutdown. The NRC never took real responsibility for its inexplicable blunder. The President's Commission:

That it was a groundless fear, an unfortunate error, never penetrated the public consciousness afterward, in part because the NRC made no effort to inform the public that it had erred.\cite{kemeny-1979}[p 126]

The NRC's mendacity, bumbling incompetence, unforgivable (not unfortunate) miscalculation, and failure to publicly admit this falsehood turned a very expensive industrial casualty into a traumatic experience that seared the thinking of many Americans. So what happened? As we shall see in a future post, all the wrong lessons were learned, while very few of the right lessons were. The NRC emerged stronger than ever, illustrating the First Rule of Bureaucratic Expansion: screw up and you get bigger. In a truly competitive market, screw up and you disappear.

1. The amount of core damage was not known at the time. Some of the misconceptions that developed later were exacerbated by the plant's owner, Metropolitan Edison's (MetEd) insistence that there had been little or no core damage, despite strong hints that there had been. As evidence mounted over the next few days that the core damage was much worse than MetEd claimed, it left the impression that ``things were getting worse". In fact, essentially all the core damage, occurred in the first 12 hours.

MetEd was not the only entity that grossly underestimated the amount of core damage. On Thursday, the 29th, testifying before Congress, NRC Chairman Hendrie was asked how close did we come to a meltdown? Hendrie replies ``Nowhere near."\cite{gray-1982}[p 184] The nuclear establishment's party line was a meltdown was not a credible event.

2. The NRC claimed, if the bubble expanded too far, it would re-uncover the top of the core. The NRC was either lying or inexcusably ignorant. The design of the B&W reactor made that impossible.

The B&W reactor pressure vessel (RPV) had a ring of check valves in the top of the RPV between the interior core hot side, and the incoming exterior cold annulus. If pressure on the core side exceeded the pressure in the cold annulus by 3.5 inches of water, these check valves would open, and the hydrogen would be vented to the cold leg rather than pushing down on the core water.

3. But it is consistent with the NRC's worstcaseitis. This is a mental disease where the patient conjures up an unrealistic to impossible worst case, in this case, no recombination of the hydrogen and oxygen; and then bases policy and regulation on something that can't happen. See also Double Ended Guillotine Break.

4. Thornburgh is the hero here. On multiple occasions, he refused to take the politically safe course of ordering a full scale evacuation, despite NRC's expert recommendations. If it were not for the governor, TMI could have turned into a Fukushima-like debacle in terms of evacuation related deaths and morbidity. Astonishing common sense and guts for a politician.

After Fukushima, NRC Chairman Jaczko said the NRC would have evacuated everybody within 50 miles of the plant. If the NRC is still around when the next American release happens, our only hope is another Governor Thornburgh.