Breaker Failure via Separate Relay 9

[Mbrooke]

Why do some utilities use a totally stand alone relay for breaker failure and reclosing? (Point on wave switching not required) What advantage if any is there? I've always thought it best to just let one relay (with a single back-up of course) do everything- makes more sense.

 

[Mbrooke]

I'm right there with you, comms exits to simply speed things up in my view, but for some POCOs that is just not the case. Con Ed doesn't even have Zone 3 enabled in a lot of their transmission relays. I don't agree with it, but its what they do.

In any case I think an over trip is acceptable risk vs a common mode failure. Systems can tolerate inadvertent trips of a few elements better than protracted faults.

 

[bacon4life]

David- Is there detailed report available about Astoria? From the NERC Lessons Learned document I had not realized there were not settings designed for in-zone faults. I thought the LL document implied that the local zone 2 protection did not trip because the fault changed types several times (LG -> 3PH -> 2LG -> LG). Is it a common thing to simulate fault evolutions during settings development?

Mbrooke-Although I agree a single inadvertent trip is generally better than a single protracted fault, I have observed many more inadvertent trips than failure to trips. I think standardizing on a single relay model greatly reduces the likelihood of an inadvertent trip and only very slightly increases the risk for failing to trip.

To avoid issues with two reclosing relays fighting, we use identical protection settings in both relays but only program the primary relay with automation/reclosing functionality. On the rare occasion the primary relay is out of service, the loss of automatic reclosing is a very small impact to reliability.

 

[davidbeach]

The NERC Lessons Learned is the only publicly available document I'm aware of. Maybe I read too much into it, but if the fault lasted over 4 minutes and it made that progression then it was a 3PH fault for many, many seconds. From what I've seen, every arcing fault starts LG and rapidly becomes something else, in this case it went LG-2LG-3LG (3PH) probably in less than a second. It then burnt until a phase burnt out, then another phase burnt out, and there was a sustained LG fault. The remote end didn't open until sometime after it became a stable LG fault. So, one can conclude, perhaps reading between the lines, that the remote end could see ground faults at the fault location but not phase-phase (including 3PH). Sure, the reactor is a complication, but even a backup phase distance element at the remote end that tripped at 2 (or even 5) seconds, could be a last grasp effort to limit the damage. Don't know if the reactor has a bypass breaker/circuit switcher, if there is it isn't shown or mentioned. If there isn't that reactor is simply part of the line impedance. Zone 2 at the remote end would need a long delay as the other lines out of the faulted station would be very short in comparison; the third terminal would also be way over-reached, needing time to coordinate with other zone 2s. Even going triple or quadruple Z2 time at the end that failed to trip would have made a huge difference.

I’ll see your silver lining and raise you two black clouds. - Protection Operations

 

[davidbeach]

As to the relative risks of over tripping and under tripping - under tripping accounts for about 4% of all misoperations and over tripping accounts for the remaining 96%. As an industry we have a little problem with failure to trip and a huge problem with being trip happy. I think we (industry wide) could do a number of things to reduce the over tripping without increasing the under tripping, and drastically reducing the number of relays and schemes in use would be a significant first step.

I’ll see your silver lining and raise you two black clouds. - Protection Operations

 

[Mbrooke]

My understanding is that the reactor does not have a bypass, and from what I was told who attended this conference:

https://events.vtools.ieee.org/m/198911

-the distance elements were set not to reach past the reactor or at least that was the intent. I do not remember as to what exactly caused the other end to open

From what I was told local breakers did open opened and generation at Astoria tripped off line.


While I hear you- respectfully consider that a sustained 345kv, 500kv or 765kv fault can cause an expansive blackout... think PJM's system as an example... At least thats how I like to deafened being trip happy.

FWIW- the third terminal is a transformer.

 

[davidbeach]

While the western interconnect is too trip happy, the eastern is far worse. The west tends toward the security end of things using POTT while the east seems enamored DCB type schemes. Maybe the more compact eastern interconnect needs to be trip happy, but when vast spaces have few lines we really don't want too many of them to trip.

Every fault will clear, I prefer to get mine cleared before anybody gets a camera turned on. As I've said before, I expect every line to have a solid foundation of step distance before any consideration is given to comm aided tripping. Make it right, then make it faster. ConEd clearly never made it right first and simply relied on making it faster.

I’ll see your silver lining and raise you two black clouds. - Protection Operations

 

[rcw retired EE]

Great discussion gentlemen! Lots of food for thought. On the Astoria "BLue Sky" failure, I was involved with the design/build construction of the Astoria-1 and Astoria-II independent power plants tying into the ConEd system. Plant 1 tied into that bus that had the spectacular failure. Plant II tied in at a new 345kV GIS substation we built and turned over to ConEd. We hired a good NY State Consulting firm to do the relaying design and commissioning.

That New Years when I saw the Astoria area switchyards on the late news, I was spooked. I wanted to call the plants and ask what happened. But I had to get my info from this forum. All the while I was wondering if my designs had created issues. Later, I learned many details that I can't share due to NDA's. Astoria 1 tripped off line and lost auxiliary power. Astoria II was not affected and continued full generation, supplying the fault through the 345-138(?)kV transformer(s). I doubt the many other generators at the Astoria facility were running. I was relieved that none of my relaying exacerbated or contributed to the issue.

I started in relaying with WWII vintage induction disk relays in the relay shop at Washington Water Power (now Avista) as a 1971 summer intern. (Summer help and summer not). EM relays were relatively easy to set and understand. A new GE relay using solid state components came in and I was assigned to help the senior tech, Andy, check it out. He was retiring at the end of that summer. His degree was from the WSU EE Department when it was still Washington State College in the 1930's. After two days of playing with it, he had taught me a lot and had tested the distance settings using resistor banks, variacs, phase shifters, timers, and a lot of meters and wires on the bench and floor. Another tech came by and asked Andy if he had that solid state figured out yet? His response was "Not really, I'm still trying to understand that new-fangled AC power."

I'm retiring in four days and feel a lot like Andy. I'm still trying to figure out AC power. I've designed a few gigawatts of power plants and have set Basler, GEC, ABB, Alstom, Beckwith, GE, Westinghouse, Siemens, SEL, Schneider, and a bunch of other brands including some Chinese relays with no manuals. I know a little bit about each but am a master of none. I totally agree with David (who used to help me with those BE relays) that you will spend a lot of time and money learning a new relay's programming language and operation. (Examples: does pu mean per unit of the machine rating, the CT ratio, the 1A/5A relay input or?? Is the setting in secondary amps or per unit? Why do the GE and SEL relays when programmed with the same IEEE standard curve have a 10:1 ratio in operating times?) If I had access to a relay and a test set we could answer some of those issues before they tripped the plant during commissioning. BTW, I don't like Breaker Failure Relays, I have tripped plants with BFI signals, more than once.

I got to do relaying about once a year or whenever a client had a problem. That meant a relearning curve every year. Having SEL in our backyard was a big help. We took a class from Lew Blackburn, who lived near our office. He would gladly offer free advice when we asked. I worry that there are still some issues out there I never knew about due to my cursory experience. I believe at least one of our forum members got to correct some of my issues over the years. That's the frustration of always building for a developer (first cost rules) and never getting to see the facilities after commissioning or find out what problems arose over time.

That is why this forum is so helpful, we can learn from each other, or I should say, I've learned a lot from your posts and questions.

Sorry I hijacked this thread, but I just wanted to wish you all good luck and Happy New Year and encourage you to keep the lights on, literally. And thank you for the help, advice and support over the years that helped me achieve some of my goals. I might do some consulting, but I'll probably just sit on the sidelines and watch you gentlemen argue/discuss and teach the younger generation.

One last piece of advice. Do not use your real name. It makes it too easy for a client to find your posts and use them against you during arbitration hearings. That is why I backed out for a couple years. Also, I enjoyed following the forum so much, my productivity was affected. Or maybe I was just getting old and slow.

Bottom line, use breaker failure logic in the relay and use redundant relays. As David said, it keeps it simple and prevents those Breaker Failure Initiate (BFI) problems.

Thanks
rcwilson

 

[Mbrooke]

@rcw retired EE: Star given, epic post that was chicken soup for my soul! Love reading stuff like this.


I want to ask and perhaps this should be a separate thread, but are there any utilities out there that did not have BF relaying in the past? Was it purely determined by cost? I don't know much about the history of BF relaying or its theory.

 

[cranky108]

In the electromechanical world, breaker failure relays, had at the simplest form two relay cases (maybe more than one aux. relay per case). So it was cost and space on the panel.

I do like hearing comments from others, and thank you.

But I still like having the breaker failure/sync-check/reclosing in a different relay than the line or transformer or bus protection. It is a matter of reducing complexity in the relay logic.
A one relay does everything is just very complex, and when you have two lines next to each other in a BAAH scheme (I have no idea why planners do that) which set of relays does what for the center breaker is confusing.
Besides, the breaker failure relay is a good location for the closing push buttons. We are going away from the control switch, because of problems with space for lights, and smart nameplates.
We still use switches for the 43 but without lights.

We do have one sub with 421 relays that do everything, including SCADA, and everyone agrees they hate it, because it is too complicated.

 

[davidbeach]

When we first introduced the 421 I spent a lot of time in the field with the Techs helping them understand the relay and expand the testing routines to cover everything the relay was doing. Success was several months later when we were testing revisions to a mid-90's installation with two 321s, a 279H, and a couple of breaker failure relays and the Techs started talking about how much simpler and how much more information was available from the 421s. It was a steep climb, but today nobody could imagine going back to the bad old days.

We have not found any reason to treat a BAAH center breaker differently if there's two lines; each line does its own thing with the center breaker and each trips the other side for breaker failure. If one line has the center breaker open and the second line trips the second line will not be attempting to reclose the breaker. Use the bus breaker as the lead breaker for the reclosing. If you're really concerned about it, double the open interval for the center breaker, so that if the second line trips just as the first is reclosing, the first line won't close the center breaker until the second line is about to close its bus breaker.

I’ll see your silver lining and raise you two black clouds. - Protection Operations

 

[Mbrooke]

Alright, I'm glad we are having this debate. What about single breaker schemes such as this?

https://www.eles.si/Portals/EN/Document/Enopolna shema_ANG - 12.07.2019.pdf

Would one or several relays be easier for testing, commissioning and life cycle cost?

 

[rcw retired EE]

Mbrooke- My 1956 edition of Mason's "The Art & Science of Protective Relaying" only mentions a breaker's failure to trip should be protected by backup relaying and implies this is the upstream overcurrent protection or the Zone 2 protection of a remote relay.

Blackburn's "Protective Relaying Principles and Applications" 1987, Section 12.26 mentions local breaker failure protection became necessary as newer EHV & UHV breakers exhibited higher failure rates and system growth required faster clearing times to maintain stability.

The Westinghouse Silent Sentinels "Applied Protective Relaying" paperback authored by Blackburn ($3.00) that we used in our WSU 1972 MSEE protective relaying class has no mention of breaker failure and just a couple comments on backup relaying.

If I had time, I'd search the IEEE Power Engineering Society transaction for "Breaker Failure Relaying" to get a historical list of Power Engineering philosophy.

BTW the Mason book states that some utilities required testing of critical bus differential CT circuits every shift for opens or shorts. It is a different world today.

Bob Wilson
Retired EE
"Don't let the smoke out!"

 

[Mbrooke]

Very interesting, I've read similar regarding breaker failure, albeit my research is rather thin.

In some parts of the world breaker failure is still not present in substations with Zone 2/3 or a reverse looking zone taking care of it. Typically there is some type of logic that stalls or prevents reclosing of the cleared ends. Read up on, and even did an evaluation where the bus coupler opens after 10 cycles then the remaining breakers clear on step distance 20-25 cycles latter.


Makes relaying simpler, but in theory also opens the door to outages. In some cases zone back up protection can be a challenge, if not impossible coordinate.

 

[cranky108]

We are having the debate on single breakers, if they need a standalone breaker failure relay. I have no problem supporting either scheme, but our testing people want to see the breaker failure stand alone.
They quote testing concerns, but we should be only testing the relays once every 12 years according to our plan, so maybe four times in their lives.
Add in testing after setting changes.

A side note to above, we also do not trip on reverse zone 3, however, we do trip on forward zone 4.
Likely sumatics with the name zone 3.

 

[bacon4life]

David- Oh, I skimmed right over the 4 minute fault duration in the intro to the Astoria NERC LL. Makes sense with that extra tidbit.

 

[Mbrooke]

If anyone is curious here is the three terminal line involved in the incident, as taken from public docs.


Two lines come from Astoria East and head toward East 179th st, both tapped feeding load transformers and generation but not connected to transmission. Both has a series reactor without bypass as Astoria East.

4-5 lines leave Astoria West. 3-4 of them feed load transformers only, two of them feed load transformers, combine, and then via single cable head to East 179th st. If I am correct this single cable has a phase angle regulator with bypass at 179th st.

I'm with others, it is unconscionable that the stand alone zone of protection does not include the reactors at Astoria East.



Bigger view:

https://files.engineering.com/getfi...b-4d29-8b49-3d27a217ca9b&file=Clipboard01.jpg

https://files.engineering.com/getfi...9-4f6b-a95d-8a023cf37f39&file=Clipboard02.jpg

https://files.engineering.com/getfi...d-423d-80a5-45c5535f8dbd&file=Clipboard03.jpg

https://files.engineering.com/getfi...a-4251-9376-cd6b1f65adb5&file=Clipboard04.jpg

https://files.engineering.com/getfi...8-411a-bb84-853a982813c8&file=Clipboard05.jpg

 

[Mbrooke]

Quick Question regarding to what was discussed earlier- I notice the 311L has an 80 bits 87L serial packet structure, and the 411L as 255 bits. This is where my knowledge about computer science runs out- but does that mean the 411 could initiate a trip faster than the 311L which worse case could lead to each relay having a different 79 state?

Typically around here is 87 primary POTT back-up, so I've always assumed there was good reason for that.

 

[cranky108]

If you are using typical 3 cycle breakers, the speed difference in tripping likely does not matter. It is a sales pitch.
I do believe the 411L is faster, but requires more communications overhead, which most companies may not think of, but is an added cost for the communications.
And yes relays with different processors will not be in sync. But the timing difference likely would not matter.