Are Regulators And The Nuclear Industry Applying The Valuable Lessons Learned From Fukushima?

Fairewinds Energy - Fukushima Daiichi Valuable Lessons

About This Video

Fairewinds Presentation to the San Clemente City Council

Fairewinds chief engineer Arnie Gundersen discusses three nuclear safety problems uncovered during the Fukushima accident that nuclear regulators and the nuclear industry wish they could ignore. Why isn't the industry designing nuclear plants to withstand the worst natural events? Why aren't nuclear regulators, governments, and citizens who live and work near a nuclear plant prepared for a nuclear accident? How much does the NRC value human life? Finally, Fairewinds' Gundersen concludes that the NRC is not implementing adequate safety changes because the NRC believes that a serious accident is impossible.

Video Transcript

Arnie Gundersen: Hi, I'm Arnie Gundersen from Fairewinds and I would like to thank the San Clemente City Council for having me make this presentation to you.  I have to apologize, my voice is a little squeeky today.  With the change of seasons here in Vermont, I appear to have picked up a virus.  But I will be O.K.

I would like to talk to you today about the lessons that Fukushima should have taught us but did not.  The first one of those is something called the design bases.  Now what that means is:  what do we expect Mother Nature can throw at us?  For instance, a plant built in California is built for a stronger earthquake than a plant built in Vermont.  A plant built in Florida is built for a stronger hurricane than a plant in upstate New York.  So that is called a design basis:  what do we think Mother Nature can throw at us?  Now in law, that comes from 10 CFR: 10 Code of Federal Regulations part 50 appendix A which is something called a general design criteria.  And general design criteria number 2 talks about design bases.

But it is interesting, it is deliberately vague.  There is no mathematical number to support the fact that an earthquake must be this strong or a hurricane wind must be this strong.  It is not in law.  The Nuclear Regulatory Commission takes that general design criteria and basically says we believe it is a good thing to build a plant for the worst thing that Mother Nature can throw at us in about a thousand years.  They go back over the historical record and they are supposed to find the worst thing that Mother Nature can do over the last thousand years of the geologic record.  Now, I do not think that has happened.  The first lesson of Fukushima, that we are not really learning, is we need to look at again, what we think is the worst thing Mother Nature can throw at us.

For example, The tsunami at Fukushima was well outside the design basis.  And so was the earthquake at Fukushima and some equipment at Unit 1 appears to have been damaged from the earthquake before the tsunami.

And two other events in the last 6 months also bump right up against the design bases.  One is the flood out in the Midwest at Fort Calhoun and the other is the earthquake on the east coast at North Anna.  Now all of these were right at or over what we thought the worst Mother Nature could do to us in a thousand years.  Now that four of these:  two earthquakes, Japan and Virginia, a flood and a tsunami, that all of them occurred in 6 months, tells me that we really have not anticipated what Mother Nature can really do.

Now let's do the math here.  The math is that, you know once in a thousand years sounds like a long time.  But really, if a nuclear plant runs for 60 years, put 60 in the numerator and in the denominator put 1,000 and you wind up with a 6% chance that any nuclear plant, over it's lifetime, will see an event as bad or worse than the design bases.  6% for San Onofre, 6% for Diablo Canyon, 6% for plants here in Vermont.  Well on top of that, there are about 60 nuclear sites.  So if you take that 6% and multiply by 60 sites, you get about 360%.

In other words, it is a near certainty that some plant in the United Sates over it's lifetime will experience an event worse than designers anticipated.  Matter of fact, more like 3 or 4 plants in the United States over the their 60 year life, will experience an event worse than the designers anticipated.  Now it is interesting though, that what the designers anticipate and what independent science anticipates are two different things.

It really boils down to cost.  The stronger you make a plant, the more costly it becomes.  So a plant in California costs more than a plant on the East Coast because earthquakes are stronger in California.  But a plant in Florida anticipates that it will get hit by a stronger hurricane than the winds you might anticipate in upstate New York.

Now outside independent experts actually have anticipated that we really have not designed for the worst case.  There were experts in Japan who said that the geologic record indicated 3 tsunamis as bad or worse than the one that hit them over a 2000 year period.  So experts in Japan, outside of the utility that owned the plant, were predicting that a tsunami could hit that was not just a 45 foot tsunami, but could even be higher, based on the record.  Those experts were ignored.  So as much as the design bases probably had been missed at least 4 times by industry experts, I think if you talk to independent experts, they will tell you that it is highly likely that a much worse event than what we have anticipated could occur.

For instance, San Onofre is designed for a one foot tsunami.  Now, on top of that San Anofre has added a margin so they can withstand about a 6' tsunami.  But on the other side of the ocean, they had a 45 foot tsunami.  I think there are experts who would say that a 6' tsunami is probably not adequate for San Anofre.

There are two things we can do to avoid this problem, neither of which is being done.  We can set a higher threshhold.  Rather than once in a thousand years, we can say once in a 10,000 year event.  Or we can listen to independent experts as opposed to industry experts when we are designing the plant.  But whatever we do on design bases, I think it is important to remember that it boils down to money.  The stronger the plant is to withstand what Mother Nature throws at us, the more likely it is to become cost prohibitive.

The second thing that I think we need to learn, and have not, has to do with emergency planning.  And within that, there are two parts.  If there is an accident, who pays?  And if there is an accident, who is in charge?  Tokyo Electric is worth about 100 billion dollars.  The event in Japan is going to cost about 250 billion dollars.  So Tokyo Electric is probably going to be driven into bankruptcy as they pay for this.  They are going to have to sell their assets and the rest is going to have to be borne by the Japanese people.

Now in the United States, it is different.  We have something called Price-Anderson.  And that limits the liability to the company that has the accident to about 10 billion dollars and the remainder, 240 billion dollars would be borne by taxpayers.  It would be the biggest industrial accident that has ever occurred within the United States.

Now, within the United States the Nuclear Regulatory Commission has allowed something to happen which actually minimizes costs.  Makes it impossible to go back at most of the utilities that own power plants.

The Nuclear Regulatory Commission has allowed them to become limited liability corporations.  Now what that means is, let's take Illinois for example:   Excelon has 17 power plants, most in Illinois.  And each individual power plant is a limited liability corporation.  So if a power plant has an accident, it has no more assets and the other power plants are not the cause of the accident, therefore they do not have to carry the bill.  The Nuclear Regulatory Commission has allowed this to happen by changing the licenses of power plants. They used to be owned by utility companies and there were assets behind them.  Now each nuclear plant is a limited liability corporation.  "Who pays?" is a really good question.

The second question is who is in charge?  In Japan I think you have noticed the confusion about who is in charge.  And I would submit to you that the Japanese are the best prepared in the event of an emergency.  They really took emergency planning seriously for years because they had earthquakes frequently.  And even now clearly, no one really knows who is in charge of cleaning up Northern Japan and who is in charge of cleaning up the site.

It is interesting, I have noticed as I have studied accidents over time that when an accident happens, the plant management recognizes really quickly that things are really bad.  At Three Mile Island the plant manager at 7:30 in the morning wanted to declare an emergency and evacuate.  Now he called the people at the home office about 150 miles away and they talked him down from that.

At Chernobyl, the same thing happened.  The plant management understood that things were really bad.  But yet the bureaucracy did not really recognize it and did not spread the word.  Of course at Fukushima we have exactly the same problem.  Plant management wanted to inject salt water.  They needed to inject salt water.  And yet higher ups in the chain of command in Tokyo told the plant manager not to.  He is a hero, he did what had to be done, despite the fact that the government told him not to.  You get this situation where the people on the ground (at the scene of the accident) know how bad things are, but yet further up the chain of command people do not make the right decisions.

In Japan the Fukushima Prefecture (like a state), had potassium iodine pills available.  What they do is they block the radiation that goes to your thyroid.   They were stocked and they were ready to be used.  But the state was prohibited from using them by the national government in Tokyo.  It was not for 7 days until the national government realized that they should release these potassium iodide pills.

Again the people on the ground (at the scene of the accident)  really recognize the severity of the problem.  But when larger organizations get involved, the time to respond lengthens and puts lives at risk.

Now in the United States the situation is probably even worse.  The Japanese understood how to do emergency planning and they still did not do it right.  Here we probably have 5 different entities that would be perhaps in charge.  First would be the utility.  Second would be the Nuclear Regulatory Commission.  Environmental Protection Agency.  FEMA.  And then also it is possible that the state could also say, "It is our job."  So we have 5 different organizations.  FEMA cannot do it, FEMA is prohibited by law to be involved for more than 30 days, something called the Stafford Act.  12:29  So they are out of the picture.

After Three Mile Island the utility was in charge briefly, and then the Nuclear Regulatory Commission came in and reported directly to the President of the United States.  Now that is not part of any law or any plan.  And I would submit to you that allowing the Nuclear Regulatory Commission to be in charge is not the best thing to do if you are concerned about the health and welfare of the people surrounding San Clemente and the San Anofre plant.

The reason is that right now there is a battle between the EPA and the NRC over the exposure to people after an accident.  The NRC wants 100 times higher exposures to the population after an accident than does the EPA.  To get an idea about what the Nuclear Regulatory Commission really plans to do after a severe accident, it is a good idea to look at a computer code they use called the MACCS2 computer program.  It is used to determine the costs and benefits to society and whether or not a utility has to implement changes to the design in order to minimize the costs to you and I.

It was designed not for a nuclear power plant accident but for a dirty bomb.  And the designer has actually renounced the program for the use the Nuclear Regulatory Commission is using.

What are some of the assumptions they have in the code?  They only look at some forms of cancer, not all, and they also do not look at other health effects caused by radiation, for instance, cesium attacks children's hearts.  And it does not cause cancer but it causes heart attacks and heart ailments.

The code does not evaluate that.  They assume that the radiation that lands on a field will be plowed under.  There is no attempt by the Nuclear Regulatory Commission to clean the fields after a nuclear accident.  They hose down the houses and let that water run into the rivers, and interestingly, if it lands on a forest, they do not plan to touch the forest.  The contamination will stay there until it decays in 300 years.  This MACCS2 program takes into account no storage of radioactive material.   There is no attempt to put radioactive material into drums and store it until it decays away.  Basically the Nuclear Regulatory Commission is assuming that it stays on the ground and in the ground, until 300 years are up and the Cesium has disappeared.

The program assumes that all the radiation stays on the ground and does not get resuspended.  A car on a dusty road throwing up dust is not included in the calculation.  Probably the most illogical assumption in the computer program is that they assume the accident lasts for two and a half hours.  Yes, two and a half hours.  Now Fukushima has been releasing radiation for 7 months but the Nuclear Regulatory Commission in their severe accident code assumes the releases occur for 2 1/2 hours.

They also assume that not much fuel is damaged, so the releases are no where near as severe as Fukushima.  They assume that the wind blows in a straight line.  As you look at the maps of contamination that came out of Fukushima, that is clearly not true either.

And last but not least, they give the owner, the plant owner, the option of paying compensation or cleaning up.  Compensation is always cheaper than clean up.  And so when the Nuclear Regulatory Commission runs this program, compensating someone for their loss is always much cheaper than cleaning up and that always turns out to be the direction the decisions are made.

So this MACCS2 program is designed to talk about costs and benefits to society.  Now even with all these assumptions which minimize the benefits to society, the MACCS2 code has actually predicted some changes should be made.  At Indian Point, it was discovered by the State of New York that 14 times the MACCS2 code said, "These changes are cost beneficial."  The state wrote to the Nuclear Regulatory Commission about this and the Nuclear Regulatory Commission responded by saying they are required by law to evaluate, to consider the changes, but even if they are cost beneficial, the Nuclear Regulatory Commission is not required to implement the changes.

I teach math at the local college here in Burlington and one of the things I teach is GIGO, garbage in, garbage out.  What that means is that the output of a computer program is only as good as the information going into it.

The Nuclear Regulatory Commission puts the lowest value on a human life of any agency in Washington.  It assumes a human life is worth 3 million dollars.  Other agencies are 5 to 9 million dollars.  So with all the assumptions I just talked about plus a low value of a human life, it is very unlikely that the Nuclear Regulatory Commission will force a utility to make modifications and it is very unlikely that you would really want them to be the agency in charge in the event of a nuclear accident.

The person who wrote the MACCS2 code is a guy named David Chanin and he has renounced it.  I wanted to share with you his own words about the code and how it is being improperly used:   "Even in 1975 the cost numbers were underestimated to a significant degree.  The underestimation is much more significant today." ... "There are quite a few things that never made sense to me, but Sandia National Labs was directed by the NRC to continue using the prior approach."  And the final quote is:  "It seems to me that the code's Quality Assurance shortcomings and the lack of input justifications are again being ignored."

This MACCS2 computer program is the key decision making tool that the Nuclear Regulatory Commission uses when they make decisions about whether a plant should be licensed for the next 20 years or when they make decisions about when a safety modification is necessary.

As I said before, GIGO, garbage in garbage out,  The code is only as good as the assumptions that go into it.  Minimize a human life or assume the cleanup is minimal, and you will justify very, very few safety modifications, which is what the Nuclear Regulatory Commission does pretty routinely.  Interestingly though, as I said in New York State, a letter to the State of New York from the NRC says that there have been 50 times when the MACCS2 computer code has determined that a safety modification would be beneficial.  And yet the NRC has ignored it even when it's code shows that a safety modification is necessary.

The real problem then lies with the Nuclear Regulatory Commission and how it implements safety modifications.  Not only does the Nuclear Regulatory Commission see no accidents, hear no accidents, speak about no accidents, but I think there is a fourth monkey too.  And that is that they believe no accident can occur.  And if that is the case, I submit to you that an accident is likely to happen because our regulator is not enforcing the regulations that are on the books.

I would like to thank the San Clemente City Council for having me tonight.  If you have any further questions or would like to study this even more, there are other videos on the Fairewinds website.

Thank you.