Tuesday, November 9, 2010

Transformer Fires? Erosion-Corrosion? Recent Events at Indian Point and Vermont Yankee

Two days ago was not the best day for Entergy. Indian Point Reactors Unit 2 was taken off-line by a transformer fire, while Vermont Yankee was shut to repair a 60-drop-a-minute leak of radioactive water within the plant. In my opinion, the best description of these incidents is in the Wall Street Journal.

Quick on the trigger, David Lochbaum of Union of Concerned Scientists was on the job about the transformer fire, while Ray Shadis of New England Coalition hopped aboard to discuss the leak. You can read their remarks in the Brattleboro Reformer.

In this case, they have made statements without knowledge. On the other hand, knowledge isn't their job. They are paid for anti-nuclear activism, not for scientific or technical credibility. Unlike science, which actually looks for causes, Lochbaum and Shadis's remarks always fit this theme: "The plant is aging and can't run anymore."

Okay. Let's look at the facts, instead.

The Transformer Fire at Indian Point

Quoting Mr. Lochbaum:

David Lochbaum,
director of the nuclear safety program for the Union of Concerned Scientists, said both problems were probably due to their age. As plants get older, they face more and more such challenges, said Lochbaum......Transformer explosions happen, on average, about a half dozen times a year at U.S. nuclear power plants, Lochbaum said.

Really? It happened because the transformer was old? The Wall Street Journal is pretty famous for digging deep, and they asked Entergy about that transformer. Here's the answer:

The transformer that failed was made by Siemens AG (SIE.XE) and installed in 2006, Nappi said.

The question then arises: Why would a new transformer fail?

Well, because it is new.

They Don't Build Them Like They Used To

Transformer fires are nothing new, but more transformer fires are happening. There's a company whose business is transformer fire prevention. (There are several companies in this market, but this is the biggest one I found by Googling.) Transformer Protection Corp.

Transformer Protection Corporation says on its website that they expect transformer fires to increase in the future because: .....This (regulatory) weakness, in addition to the globalization market, has opened the door to a disturbing drop in the quality of new transformers...Many experts anticipate that the number of failures will increase significantly in the near future, from 1% in 2001 to 2% in 2008 ..... In addition, the shorter lifetime of new transformers will sharply increase above this rate after 2008

The website also notes that there were around 800 transformer fires in the U.S. in one year. I have no idea if Lochbaum's "a half-dozen fires at nuclear plants" is correct. If it is correct, I can rephrase his statement as below.
  • Less than 1 % of the transformer fires in the U.S. were in nuclear plants last year.
  • Nuclear plants should be careful to not replace transformers unless they absolutely need to do so. The old (aging) transformers usually have better life expectancy than their replacements.
Erosion Corrosion

For years I was a member of NACE and a working corrosion engineer. Mr. Shadis's description of erosion-corrosion is accurate. It is a terrible, sometimes catastrophic form of corrosion. It breaks pipes.

In other words, erosion-corrosion rarely seeps. It busts things. It causes pipe failure. I would say "Erosion-corrosion never seeps" but I have been a corrosion engineer too long to be throwing around the word "never." Things happen, you know. Not everything goes according to form. Erosion-corrosion could make one through-wall pit, under some complex circumstances, a single pit that leaks. It would be unlikely to leak at a drop-per-minute rate, however. Also, erosion-corrosion is rarely seen in straight flows of steel pipes. It is far more common in copper, or at places where the flow in a pipe becomes turbulent.

Let's say that erosion corrosion would NOT be the first thing I would consider in a case of seepage in a straight run of pipe. Indeed, Vermont Yankee had seepage around a weld, the most common thing on earth, corrosion-wise. Mr. Shadis decided that to say that the seepage was erosion-corrosion because that was the worst thing he could think of. It was not a likely cause for the leak.

As I said, Shadis is paid to attack the plant, and that is what he does. Most of the time, frankly, he does it better than he did in this case.


Dstarr said...

I'm hearing that the Vermont Yankee leak is from a pipe weld done in 1972. So reweld it, and get on with it. Leaks happen.

Mike Mulligan said...

Why did it fail, because we don’t make transformers in American anymore!

Raymond Shadis said...

Hello Meredith,
Please take issue with the content of my comments, but do not impute intent. I did not pick the worst phenomenon of which I could think. Nor did I hop on the comment train. A reporter called to ask my take on the very little bit of information that Entergy had released. At that point it was uncertain if the leak was restricted to the plug and seal weld or if the pipe was aloso leaking in the weld area of a collar around the plug. The reporter asked what I thought might be causing the leak and I gave him a range of possibilities..including the possibility that if there was a through-wall leak down-stream of the port, flow-accelerated corrosion could be a factor. So, sue me! But look, what is the cause of the 2010 leaks(turns out now it could be more than one)? Well, maybe its a design failure in what should have been leak-proof plugs. Maybe its just shitty welding- a Q/A failure or inappropriate selection of weld material. Maybe the welds should have been stress-relieved to avoid stress corrosion cracking. Don't know and Enertgy is unlikely to be forthcoming about it. This much we can reasonably surmise: Entergy VY's extent-of-condition review following a nearly identical leak elsewhere in the FW system in January 2009 failed to catch the incipient failure which matured in in November 2010.
You've chosen a difficult task in saving VY when its owner-operators daily work to drag it down.
Raymond Shadis

Kit P said...

“where a spare transformer was on site.”

As long as we are speaking of fires in electrical equipment, there have been more than 50 home fires caused by solar systems (aka, smoke emitting diode). We have to wonder if any spare homes were on site. Fortunately, no one has been injured.

Is Lochbaum also involved in the UCS home safety program? Idiots!

Producing electricity involves many hazards but the electricity producing industry has a very good safety record. Taking PCB out of transformers did not make them any safer considering they are none to fail.

Erosion-corrosion is also a know problem in steam plants. There were several deaths in the nuclear industry when pipes failed but I think has been more than 20 years since there has been an event in the US. Reliability programs now now monitor the issue to fix piping before it fails.

The last event I know of was in New England at a coal plant just a few years ago.

Anonymous said...

I'd like to take issue with one more thing. I think we need to start making a big deal about any news source which uses the expression 'radioactive {material}' (where, in this case, {material} is water), without explaining how radioactive the material actually is, and without explaining what level of threat it presents.

To Joe Q. Public, if something is described as "leaking radioactive water", that means the water is very radioactive, at least radioactive enough to be a substantial hazard. Is that true in this case? I have NO IDEA because the news source doesn't bother to report such an important fact.

"Radioactive" isn't really a 'boolean value', but the media treats it as such. By boolean, I mean a value which is either true, or false, with a very clean 'boundary' between true and false.

'Radioactive' on the other hand, can just mean that a material puts off very very small amounts of radiation, smaller than would be considered any significant threat. It could also mean it's putting off very large amounts of radiation and will give you a lethal dose within minutes (or seconds) of exposure, but if the journalist doesn't explain that, we're left to just guess at what level of risk is present in the situation.

I *hate* that.

Meredith Angwin said...

Thank you all for your comments!

DStarr and Mike: I agree with both of you.

Raymond. I am happy that you actually gave the reporter a list of options, and sorry I assumed that the reporter quoted you completely. Quoted-out-of-context is a serious hazard of speaking in public, yet the public discourse would be poorer if people decided not to speak. As you must have noted, I did note that this wasn't your usual approach.

You also said something about how badly Entergy is running the plant. I think it is being run well and the recent number of breaker-to-breakers is my proof of this. We must simply agree to disagree on this topic. We will never convince each other, that's for sure!

Kit. Erosion-corrosion is definitely a known problem in steam plants, and I hope I didn't come across as dismissing it. There was a case in the Surry nuclear plant in 1986 which killed four people. There have been no such accidents in U S nuclear plants since. (One in Japan I think.) We have predictive computer programs and inspections to deal with this mode of corrosion. My point was simply that this is usually a catastrophic failure situation, not a situation of seeping one drop per second. Naming this type of corrosion for a minor seepage struck me as unreasonable.

Jeff. Quite right. "Radioactive" is not a boolean yes/no. I would also like to change the discourse in this regard.