High Power Circuit Breaker Operational Reliability Question

[EverGreen1231]

My supervisor asked me Friday if there were any way to determine the operating reliability of a large circuit breaker in a Substation. Now, to me (just having thought about it a little over the weekend), it would seem that breaker reliability of operation would really have to be something that would have to be tested since it relies on too many factors to reliably calculate.

The reason this came up is we (the power company I work for) give out yearly breaker operation reports to all of the towns/municipalities that we serve. We use the number of operational events there were and how many operational events that lead to lockout or events that went straight to lockout on each breaker for that cities' substations. This gives the city some idea of distribution lines that may have some too much tree growth around them, or if there's some other kind of reoccurring fault that can be fixed.

With this in mind, I'm thinking her question is more along the lines of system reliability and not breaker reliability, which is a whole different ball game. I was wondering if anyone had any decrypting articles or e-books handy regarding high power circuit breaker reliability (Jim Hardy, I'm looking at you ) : How it could be quantified; what is it, and how does it tie into system reliability? among other questions I'd like to know. I spent some time today and yesterday trying to find things on Google with no success.

Much obliged,
-Tyler Glenn.

 

[anorlunda]

There are two ways to look at it.

If failure/nonfailure of a component is binary, and if failure of a component does not cause cascading failures, then a Monte-Carlo statistical treatment using MTTF and MTTR of components is the way to do it. Failures in power distribution networks, for example, are usually treated that way.

On the other hand, failures of a simple component can on rare occasions cause transients that cascade into large-scale blackouts. The most famous example is a simple relay at the Sir Adam Beck Dam which triggered the entire Great Northeast Blackout of 11/9/1965. To predict those kind of consequences, one postulates a failure, then uses dynamic simulaiton of the entire power grid to predict the consequences. Utilities have system planning departments that do such simulations day-in day-out for the power transmission networkds.

Note the distinction between power distribution and power transmission.

 

[EverGreen1231]

There are two ways to look at it.

If failure/nonfailure of a component is binary, and if failure of a component does not cause cascading failures, then a Monte-Carlo statistical treatment using MTTF and MTTR of components is the way to do it. Failures in power distribution networks, for example, are usually treated that way.

On the other hand, failures of a simple component can on rare occasions cause transients that cascade into large-scale blackouts. The most famous example is a simple relay at the Sir Adam Beck Dam which triggered the entire Great Northeast Blackout of 11/9/1965. To predict those kind of consequences, one postulates a failure, then uses dynamic simulaiton of the entire power grid to predict the consequences. Utilities have system planning departments that do such simulations day-in day-out for the power transmission networkds.

Note the distinction between power distribution and power transmission.

Yes, the component is binary in nature. It either operates, or not.

The Monte-Carlo statistical treatment sounds like it would work since the distribution system of the cities are radial and don't tie into each other accept at the substation/transmission level (which is where the planning department takes over). Do you have any documentation shortly on hand that could lead me to a method of analysis? I ask because I've honestly never heard of this before (that I remember).

Thank you for your answer!

 

[anorlunda]

Yes, the component is binary in nature. It either operates, or not.

The Monte-Carlo statistical treatment sounds like it would work since the distribution system of the cities are radial and don't tie into each other accept at the substation/transmission level (which is where the planning department takes over). Do you have any documentation shortly on hand that could lead me to a method of analysis? I ask because I've honestly never heard of this before (that I remember).

Thank you for your answer!

https://en.wikipedia.org/wiki/Monte_Carlo_method

Don't forget that even a simple component like a breaker can have several failure modes. Fails to open, fails to close, opens but fails to quench the current, internal short circuit, uncommanded open/close ...

Good luck.

 

[EverGreen1231]

https://en.wikipedia.org/wiki/Monte_Carlo_method

Don't forget that even a simple component like a breaker can have several failure modes. Fails to open, fails to close, opens but fails to quench the current, internal short circuit, uncommanded open/close ...

Good luck.

Thank you. If you don't mind my asking...do you have experience in power systems?

 

[anorlunda]

Only a little; 45 years. But I've been retired for 10 years now so a lot of my experience is out-of-date.

 

[EverGreen1231]

Only a little; 45 years. But I've been retired for 10 years now so a lot of my experience is out-of-date.

I ask because I'm entering a fledgling career and I really enjoy it, and I always like getting advice from others that've been where I am now . I work a lot with Protection and Operations & Maintenance. I don't know what kind of future may be in this industry for me, but I like the close knit community there seems to be among those involved with it.

 

[jim hardy]

My supervisor asked me Friday if there were any way to determine the operating reliability of a large circuit breaker in a Substation.

That's vague .
You have done the right thing in trying to figure out what it is they really want to know.

I was in a power plant, distribution was another group with whom i only fraternized.

Taking their question literally, one would look for statistics on the type of breakers involved.
My utility had a quality group to whom collection of such statistics fell.

EPRI was formed in the 70's to study such things for the industry.
A quick search on EPRI circuit breaker reliability returns a ton of hits.
EPRI is sort of stingy with their information though, if your utility is not a member they won't let you access a lot of it.

http://www.epri.com/search/Pages/results.aspx?k=Cost%20Effective%20Circuit%20Breaker%20Monitoring%20and%20Management

I'm not above contacting authors of industry journal articles that seem relevant to my inquiry
http://tdworld.com/substations/breaker-maintenance-more-tool

that's how i learned about co-ordinating fault protection for inverters.

Sorry i can't help you more - just it's not an area where I'm strong.
Those EPRI links even had studies on ancient GE MagneBlast medium voltage switchgear, which is what we had in the plant. But T&D is a different world.

Train your search engine - search on keywords from articles that look close to what you want.
Google will try to please , and it pays attention to where you spend time. Toss it a few bones.

it would seem that breaker reliability of operation would really have to be something that would have to be tested

The only real test is to put it out in the real world and gather statistics. Rest assured somebody has gathered them.

But as you discerned, the question as presented to you may have suffered in translation en route.
The issue may well be continuity of power to municipal feeders.

And i recognize that as "new kid on the block" you need to question with certain, well, demeanor..
Get clarification,
perhaps by rephrasing the question as
"So they're asking about probability that our breaker will trip when it needs to ?
I wonder if they meant to ask instead about probability of power outages on particular feeders? Seems to me that'd make more sense, from their shoes. And since our breakers have operation counters(i assume they do-jh) , it should be easy enough to tabulate an estimated cycles per year for them. "

 

[Windadct]

Determining the reliability of ONE breaker? -- Will seem remarkably high, however the cost of a mis-operation is also high. For a population of CBs, due to the variety of designs and ages - I would catalog the population and discuss with each of the OEMs as well. For the breakers and the relaying systems. IMO -- human error (including neglect) leads to the most failures, so the operational and maintenance of the devices also should receive serious scrutiny.

I did Circuit Breaker Maintenance in the field for a number of years, simple exercising seemed to be the most effective, as well as through inspection, and BASIC cleaning - as soon as part begin to be dismantled and re assembled the skill / training of the FSEs and Technicians becomes important.

 

[EverGreen1231]

,

That's vague .
You have done the right thing in trying to figure out what it is they really want to know.

She's asked similar of me before. She asked me to "figure out a way to analyze the health of our transformers." After reading nearly every article/whitepaper I could find on the subject and an IEEE C57 later and I know more about transformer health than I ever thought there was to learn and have still only scratched the surface. But I enjoy learning and I'm decent at producing something useful without much direction.

I was in a power plant, distribution was another group with whom i only fraternized.

Taking their question literally, one would look for statistics on the type of breakers involved.
My utility had a quality group to whom collection of such statistics fell.

EPRI was formed in the 70's to study such things for the industry.
A quick search on EPRI circuit breaker reliability returns a ton of hits.
EPRI is sort of stingy with their information though, if your utility is not a member they won't let you access a lot of it.

http://www.epri.com/search/Pages/results.aspx?k=Cost%20Effective%20Circuit%20Breaker%20Monitoring%20and%20Management

I'm not above contacting authors of industry journal articles that seem relevant to my inquiry
http://tdworld.com/substations/breaker-maintenance-more-tool

that's how i learned about co-ordinating fault protection for inverters.

Sorry i can't help you more - just it's not an area where I'm strong.
Those EPRI links even had studies on ancient GE MagneBlast medium voltage switchgear, which is what we had in the plant. But T&D is a different world.

Train your search engine - search on keywords from articles that look close to what you want.
Google will try to please , and it pays attention to where you spend time. Toss it a few bones."

Thank you for your links, sir. I'm going to read-up and see what can be used. I'm like you in that I'm not above contacting folks that seem to know more about a subject than I, or more to the point that they can write a paper on the subject.

Google pretty much knows what I look for. I think a lot of the reason I sometimes don't find anything I deem useful is because I don't know what can be used. That's when I come here to maybe get a little direction/clarity on the priciples... I have seldom been disappointed.

The only real test is to put it out in the real world and gather statistics. Rest assured somebody has gathered them.

But as you discerned, the question as presented to you may have suffered in translation en route.
The issue may well be continuity of power to municipal feeders."

I'm thinking so as well. She was asked by her supervisor which then got passed on, in a slightly different form, to me.

And i recognize that as "new kid on the block" you need to question with certain, well, demeanor..
Get clarification,
perhaps by rephrasing the question as
"So they're asking about probability that our breaker will trip when it needs to ?
I wonder if they meant to ask instead about probability of power outages on particular feeders? Seems to me that'd make more sense, from their shoes. And since our breakers have operation counters(i assume they do-jh) , it should be easy enough to tabulate an estimated cycles per year for them. "

I do have to tread lightly. There's nothing worse than a "young buck" getting too 'ambitious' in his questioning...understandably so. My understanding from my supervisor is that her supervisor asked about finding a "creative" way to look at the breaker operation counts (I suppose he really meant to say 'meaningful', but I could be wrong). I'm thinking along the same lines as you that something got lost in translation. I would try and get clarity if either were in the office. I suppose the best course of action would be to chase a few of the fatter rabbits around until they return sometime next week.

 

[Jeff Rosenbury]

This is likely a naive question, but do circuit breakers wear with age? Perhaps another way of asking this is, how much correlation is there between the age of a unit and failure rate?

 

[EverGreen1231]

This is likely a naive question, but do circuit breakers wear with age? Perhaps another way of asking this is, how much correlation is there between the age of a unit and failure rate?

Yes, they age. I would say in slightly different ways depending on what type it is (Vacuum, Gas, oil). I don't know that there's a correlation to breaker age and reliability of said breaker's operation (not failing to close, not failing to open, not failing in quenching the arc, etc.); it'd definitely be something worth looking into...my intuition says that it certainly must be so, but I'm not sure as to the rate that reliability diminishes as said aging occurs.

 

[EverGreen1231]

Determining the reliability of ONE breaker? -- Will seem remarkably high, however the cost of a mis-operation is also high. For a population of CBs, due to the variety of designs and ages - I would catalog the population and discuss with each of the OEMs as well. For the breakers and the relaying systems. IMO -- human error (including neglect) leads to the most failures, so the operational and maintenance of the devices also should receive serious scrutiny.

I did Circuit Breaker Maintenance in the field for a number of years, simple exercising seemed to be the most effective, as well as through inspection, and BASIC cleaning - as soon as part begin to be dismantled and re assembled the skill / training of the FSEs and Technicians becomes important.

I also think that human error has a lot to do with it. There could also be a misplaced trip signal from a rogue relay (the microprocessor types sometimes do that). I think, for my purposes, outside stimuli can be ignored, or, if not ignored, a safe margin of "error" given for the consideration of outside equipment/personnel mistakes.

 

[Windadct]

My point on human error is - if the OEM provides a maintenance schedule and the procedures are not followed properly ( they could be too difficult, require special tools that are not available - etc) will cause more failures then the equipment its self.
Examples - Large OCBs are typically closed with an Air Cylinder - if the compressor is not maintained or the air reservoir fills with water, common in humid climates - then the breaker may not fully close. - This is really human error -as the equipment was not properly maintained.
Most systems operate with Battery supplied DC - again batteries require proper maintenance.
A properly applied, maintained breaker will have an exceptionally low failure rate. A "rogue relay" should be statistically trapped by the OEM. In some of the 1st and 2nd generation microprocessor types - there were bugs, but they should all be replaced or upgraded today. A relay mis-operation is too serious.

So - to the OP - Who Made it, was it selected properly, was it installed properly, when was the last time it was exercised, maintained and who did the work. When was the last relay audit and coordination study done, when were the relays maintained and calibration checked.

Determining health of transformer is more straightforward - IMO, and health vs operating reliability are kind of different questions. Both can be addressed by a NETA contractor.

 

[jim hardy]

My understanding from my supervisor is that her supervisor asked about finding a "creative" way to look at the breaker operation counts

hmmm

In what form is operation counts available ? Hopefully tabulated by date/time ?

Put 'em in a spreadsheet and plot against weather, load, time of day, construction projects, substation maintenance, whatever else you can think of.

Weibull ?

 

[EverGreen1231]

hmmm

In what form is operation counts available ? Hopefully tabulated by date/time ?

Put 'em in a spreadsheet and plot against weather, load, time of day, construction projects, substation maintenance, whatever else you can think of.

Yes, SCADA keeps up with all that info and puts it into the Oracle database. There's a spreadsheet already prepared for those numbers...thankfully.

I think this and maybe the Monte Carlo Method mentioned a few posts ago will do the trick.

Thank y'all for your posts, insights, and advice.

 

[EverGreen1231]

My point on human error is - if the OEM provides a maintenance schedule and the procedures are not followed properly ( they could be too difficult, require special tools that are not available - etc) will cause more failures then the equipment its self.
Examples - Large OCBs are typically closed with an Air Cylinder - if the compressor is not maintained or the air reservoir fills with water, common in humid climates - then the breaker may not fully close. - This is really human error -as the equipment was not properly maintained.
Most systems operate with Battery supplied DC - again batteries require proper maintenance.
A properly applied, maintained breaker will have an exceptionally low failure rate. A "rogue relay" should be statistically trapped by the OEM. In some of the 1st and 2nd generation microprocessor types - there were bugs, but they should all be replaced or upgraded today. A relay mis-operation is too serious.

So - to the OP - Who Made it, was it selected properly, was it installed properly, when was the last time it was exercised, maintained and who did the work. When was the last relay audit and coordination study done, when were the relays maintained and calibration checked. .

I see your point...but, I don't know that the causation behind the possibility of the kind of misoperation/failure is important (OEM procedures and such...) to what my supervisor was looking for. Now, to me, if I were one of the "higher-ups", I'd want to know what the most likely cause of the failure was so that we could alter procedures or take some other course of action (if any is necessary). If they indicate that the causation behind the failure is not important, than the method mentioned in the first reply to this thread, I think, would be viable.

Determining health of transformer is more straightforward - IMO, and health vs operating reliability are kind of different questions. Both can be addressed by a NETA contractor.

Transformer health is much easier to determine (IMO). Once you know the various types of high voltage transformers there are and, basically, how they're built and what each part is there to do, than it becomes 'easier.' I suppose health and operating reliability are somewhat different, but I would say you could derive one from the other. If a breaker is 'healthy' it has a very low chance of failure, and reliability is good; if it's not, it has a higher chance of failure, making operational reliably fall...but you know this, I saying it, mainly, for my own benefit.

 

[anorlunda]

System reliability may or may not be part of the OP's responsibility. But for the benefit of others who may read this thread, let me explain better the relationship between system reliability and component reliability. See also post #2.

Consider a simplified distribution loop as shown below. The substation is the interface between the distribution system and the transmission system. Most distribution circuits are arranged as loops, that can feed power clockwise or counterclockwise, thus providing two paths to most loads. The cross hatched rectangles are components (breakers, transformers, switches, fuses, overhead line sections, underground cables, shunt capacitors, whatever). Let's assume that we know of a mean time between failues MTBF and mean time to repair MTTR, for each failure mode (say short circuit, open circuit, whatever) for each component.

The arrows are places where one or more customer loads can be connected.

The first thing to observe is that the chance of power interruption is different for each node in the circuit. In general, higher reliability closer to the substation. So we need to calculate interruption frequency and duration averaged over all the loads.

Next, note that the consequences of failure of a component depends on its location in the circuit. In the simplified case above, there is no load that will be interrupted by a single failure, it would take two or more simultaneous failures. Simulaneous failure of the two components closest to the substation takes out the whole loop. Simulaneous failure of the two components furthest from the substations takes out only one load.

In real life, single failures may interrupt a few customers while leaving others undisturbed. Think of a branch taking down the wire from the pole to your house (withough causing a short circuit). That single failure will take your house out, but nobody else. Also in real life, the typical loop will have thousands of components and topologies more complicated than a single loop. Also some loads more critical than other loads.

Tree branches taking down lines is the most frequent bad event. Squirrels crawling on components is also very significant.

Next, we use the Monte Carlo statistical method to calculate the system average frequency and duration of interrruptions, based on the component MTBF and MTTR values, and from the component placements in the circuit. That is the system reliability calculation.

Finally, we can calculate the contribution of each component to system reliability, by first calculating system reliability with that component in the circuit and calculating again with the component removed. The difference in the two answers is the contribution of that single component to overall system reliability; it is a meaningful number. The same method can be used to calculate how much value there is to doubling or halving the MTTR of a single component.

So, the question in the OP is quite doable. The contribution of a single breaker to system reliability can be calculated, and that number is meaningful. Whether or not that should be the reponsibility of the poster is not an engineering question.

Sorry if I went overboard with this long post. It is both overly complicated and overly simplified. I couldn't resist.

 

[EverGreen1231]

System reliability may or may not be part of the OP's responsibility. But for the benefit of others who may read this thread, let me explain better the relationship between system reliability and component reliability. See also post #2.

Consider a simplified distribution loop as shown below. The substation is the interface between the distribution system and the transmission system. Most distribution circuits are arranged as loops, that can feed power clockwise or counterclockwise, thus providing two paths to most loads. The cross hatched rectangles are components (breakers, transformers, switches, fuses, overhead line sections, underground cables, shunt capacitors, whatever). Let's assume that we know of a mean time between failues MTBF and mean time to repair MTTR, for each failure mode (say short circuit, open circuit, whatever) for each component.

The arrows are places where one or more customer loads can be connected.

The first thing to observe is that the chance of power interruption is different for each node in the circuit. In general, higher reliability closer to the substation. So we need to calculate interruption frequency and duration averaged over all the loads.

Next, note that the consequences of failure of a component depends on its location in the circuit. In the simplified case above, there is no load that will be interrupted by a single failure, it would take two or more simultaneous failures. Simulaneous failure of the two components closest to the substation takes out the whole loop. Simulaneous failure of the two components furthest from the substations takes out only one load.

In real life, single failures may interrupt a few customers while leaving others undisturbed. Think of a branch taking down the wire from the pole to your house (withough causing a short circuit). That single failure will take your house out, but nobody else. Also in real life, the typical loop will have thousands of components and topologies more complicated than a single loop. Also some loads more critical than other loads.

Tree branches taking down lines is the most frequent bad event. Squirrels crawling on components is also very significant.

Next, we use the Monte Carlo statistical method to calculate the system average frequency and duration of interrruptions, based on the component MTBF and MTTR values, and from the component placements in the circuit. That is the system reliability calculation.

Finally, we can calculate the contribution of each component to system reliability, by first calculating system reliability with that component in the circuit and calculating again with the component removed. The difference in the two answers is the contribution of that single component to overall system reliability; it is a meaningful number. The same method can be used to calculate how much value there is to doubling or halving the MTTR of a single component.

So, the question in the OP is quite doable. The contribution of a single breaker to system reliability can be calculated, and that number is meaningful. Whether or not that should be the reponsibility of the poster is not an engineering question.

Sorry if I went overboard with this long post. It is both overly complicated and overly simplified. I couldn't resist.

Please, no apologies. Your posts are insightful, interesting, and (most importantly) readable to us mere mortals . I'm approaching the conclusion that this is more a question for the gentlemen in planning, but I'd be interested in figuring out how to go about doing it. I'm learning about Monte Carlo methods and trying to relate to how they could be used here in my spare time...it could take me a bit but that's only because I have to wrestle with something a while before I come to an understanding.

 

[anorlunda]

I'm learning about Monte Carlo methods

Google "distribution system reliability evaluation" I found Distribution system reliability evaluation using the Monte Carlo simulation method