Last January, the Byron Station, Unit 2, automatically shut down its nuclear reactor because of an electrical power problem. But the way the power was lost – only one phase of three-phase alternating current was lost – in addition to a design problem caused the diesel-driven back-up generators to not start up automatically.
While operators were able to get the back-up generators started manually and the plant was shut down safely, this was clearly a problem the NRC has to look into. Since power can be lost to a plant for any number of reasons – including a thunderstorm or other natural events – back-up power is very important to keep vital safety systems working at the plants.
The NRC quickly issued an “information notice” to let other nuclear power plant operators know about the problem. The NRC then, some months later, followed up with a bulletin to all licensees. This bulletin required all 104 reactors and the four licensees that had submitted new reactor designs to provide the NRC with information about their electrical protection systems. The NRC staff will use the information provided by the licensees to determine if further regulatory action is needed.
A bulletin and an information notice are two types of NRC “generic communications.” Along with generic letters and regulatory issue summaries, generic communications provide a method to communicate, primarily with licensees, about issues that might affect all of them, even if the original event – as in this case – is specific to just one plant.
The term “generic” might seem odd in this context – most people probably think of generic in terms of medicine or consumer products. But for the NRC, the term refers to important ways we communicate generally about issues that might affect more than one site.
NRC generic communications can be viewed on the NRC’s public website.
Andrea Russell Project Manager Generic Communications and Power Uprate Branch
U.S. NRC Blog 
Nuclear guy, you seem very educated regarding nuclear energy. I’m thinking SRO possibly ? I would like to follow your interactions,do you have a blog or some other public forum ?
With regards to San Onofre, they put the heat exchangers in under the 50.59 rule, which allows them to make replacements that are not like-for-like provided they meet specific criteria and are not affecting their operating license. Now they may have misapplied the rule (and if that’s the case I would expect a violation to come out of this in the future), but plants ARE allowed to make major changes, even ones which are not like-for-like, if they can pass the 50.59 screening or evaluation.
Technically you are correct, in practice though I think the NRC needs to be a stronger regulator. Not allowing a situation where it is better to ask for forgiveness than to ask for permission. LIke at San Onofre….they should be driven out of business for lying about the so called like for like changeout of the exchangers, when in fact the new exchangers were far different and in fact became a huge problem very quickly. They need stiff penalties so that those corporate quarterly profit pumping bean counters think three time before trying a stunt like that.
http://www.thefreedictionary.com/generic
“Relating to or descriptive of an entire group or class; general.”
Generic does not mean “no nevermind”. Any letter from the NRC, unless otherwise stated, also does not mean “no nevermind”. Generic means it has applicability to a large group or class. If people somehow think generic just means its a “no nevermind” they are applying an assumption to their interpretation of the word.
This could be tested in a simulator though. I know my plant did simulator scenarios as compensatory actions to ensure operators understood how to identify this issue if it were to happen and what the proper responses were.
Fresh, I apologize I was a little unclear. When this event first happened, I was very intimately involved with the initial understanding to see how it would apply to our plant.
I agree a phase dropping is a huge problem. To understand why protectives didnt automatically isolate the power source and auto-transfer to the main power source, you need to understand the main two protections. The first is loss of power, which is a set of relays in a 2 out of 2 logic. These relays have to look across 2 phases, and will typically look at A-B, B-C. or C-A. If you have a plant that, say, has relays across A-B, and relays across B-C, and you have a loss of the C phase (which is what happened at Byron Unit 2), you would only get a protective signal on the B-C relay, only giving you a 1 out of 2 and not causing the protection to kick in. This is per design, as these relays are not designed for single phase detection. If you had this happen on the B phase in this scheme you would get the 2 out of 2 protection.
The second protection is differential current. In this unusual case, the line side remained suspended in the air as an open circuit. The load side fell to the ground without faulting. That is to say, it fell down, but it didnt touch anything which would allow electricity to conduct towards ground. Because there was no ground path, not enough differential current could pass BACK through the damaged load bus bar to a ground path, and this protective didnt pick up. If there was a fault on the line side, or if the ground side faulted to a live conductor, this protective would have picked it up. You typically would count on this protective for a phase issue, but this protective cannot work without an actual fault.
The loss of phase caused reactor coolant pumps to trip offline, and an automatic reactor scram. Afterwards, the loss of main steam supply caused the loads supplied by the generator to transfer back to the reserve power source, further exacerbating the problem. 480V unit substations were tripping on thermal overloads and the like. The operators looked at their voltage indication and they were looking at the A-B relay which was showing good voltage and there was some confusion. A short time later (minute or two i think), they checked the other phases and saw degraded voltage conditions, and manually opened the reserve power connection. Once that happened the LOOP logic for the plant picked up and the diesel generator emergency sequencer kicked in. During the 8 minutes w/out power, core cooling was supplied by a diesel engine driven aux feedwater pump, and decay heat was removed for the entire cooldown using atmospheric steam dumps.
The protection scheme for open phase was designed per NRC guidelines and approved by the NRC. It was the unusual case where the load side separated from the line side without faulting that caused the problem. A month later Byron Unit 1 had the exact same failure happen, except it occured on the load side, and the plant automatically did a fast bus transfer and stayed online (which is what it should have done). This was a fringe case that was pretty unusual.
Anyways for lining oh a phase, ultimately what I meant was the meter to detect power on a specific phase. If your particular plant logic is a relay on the A-C phase, and a relay on the B-C phase, then a failure of the C phase will cause a 2 out of 2, but a failure of A or B will not, and ensuring you have meters on B and A will give you direct indication of the failure. A large part of the required compensatory actions to this issue was ensuring that the operators had reliable means of detecting this issue, and understanding the manual action to be taken.
The entire industry is vulnerable to this event, and every plant had to perform operability evaluations and implement compensatory actions. Every plant is also in process of designing and implementing industry wide fixes to this (I’m actually not sure what the fix looks like since it left my group when that happened, but I’ll probably know when it comes time to implement it during our next outage).
What happened at Byron was an unanticipated fringe case, but it is not an acceptable one. What is ironic is the news media was trying to somehow spin this into a “radioactive steam contaminating chicago” type thing, and completely ignored the fact that a real vulnerability was discovered.
I can also tell you that if we had known that this would have happened, if we had any idea we could end up in this case, at least with my plant and the fleet I work with, we would have implemented compensatory actions much sooner. The risk of core damage is not just a public health one, but a financial one for the company, and I can only speak for my experience working at a few BWRs, but the owners of plants have just as much of an interest in safety as safety also means asset protection.
Just my thoughts.
Dropping a phase is a classic and huge problem. Its a great way to burn up a motor or a whole nuke plant.
I am flabbergasted that the nuke industry has come 70 years without addressing these core and critical issue. Living on borrowed time is the phrase that comes to mind.
What is the next “admission” going to be, with a half baked “response”, only the insiders with an open mind really know the risks the industry is taking to maximize profits to the corporation that hold them close, while providing MAXIMUM campaign contributions to influence and create a greater lack of oversight.
Sorry that is pretty stinging, and probably will be censored out, but it is the truth. And of course, big oil, big union, and now even big solar are getting into the act.
Say the phases are A,B,C, how does lining up the B phase only improve anything? What if A or C drops out?
Generic in “plain english” comes with the negative connotations that I mentioned.
Generic means it applies generically to all plants and is not a specific issue. It’s not really industry specific language….that’s plain english. There’s not even an acronym here (and the nuclear industry is very notorious for its excessive use of acronyms).
Any letter or notice from the NRC cannot be ignored.
Plants had to perform operability evaluations and establish compensatory actions immediately after the incident. This includes things like having your safety power sources split between your normal and emergency transformers, making sure the voltage meters were lined up to the B phase to ensure you can manually detect loss of voltage, and pre-job briefs in case a similar event were to happen at your plant. This was required in order to maintain the offsite power system operable. This includes training on the issue as well as updating procedures.
This is really the best interim solution until a proper fix can be designed and implemented, to ensure you have actions in place that would prevent core damage should the same initiating event happen.
The sad reality is that a real life test as you propose could result in a real emergency, a real meltdown. Rarely are emergency systems given a full dress rehearsal, to the individual humans involved, that risks are just to high, and the ability to easily slough off the rest test is attractive to everyone involved.
Gerneric sounds to low level, kind of a “you can ignore this”, gets no respect like a generic box of oats on a Costco shelf. Just like in the legal industry, they cover their operations with many industry specific words, really, even latin. This is the first line of defense against the average Joe from being able to figure out what they are doing.
So if the NRC is making a general comment, or really better a general request for information, then it should be called that and not generic.
RE: “The NRC quickly issued an “information notice” to let other nuclear power plant operators know about the problem. The NRC then, some months later, followed up with a bulletin to all licensees. This bulletin required all 104 reactors and the four licensees that had submitted new reactor designs to provide the NRC with information about their electrical protection systems.”
What would be a much better “Solution” would be to have the on site inspector, simulate the above problem and then record exactly what the reactor CREW did about it. This information would provide the NRC with a much better picture of not only how operators reacted to this “new” problem but how able they were to “solve” it. Then that information could be studied by the NRC before they issued a “bulletin” some months later, along with site specific additions to their operating manuals…