Low SF6 action 8

[Mbrooke]

What are the typical choices of action when dealing with low SF6 and an SF6 cutout? Is an automatic relaying action initiated or is the decision left to operators via remote SCADA?

 

[Mbrooke]

Minus deliberate vandalism, are there any known leaks that rapidly go from low to low-low?


@DavidBeach: those breakers that raise alarm, is it because of a defect (breaker needing repair/replacement) or are such leaks "normal" during the in service life?

 

[davidbeach]

Normal leakage. I would guess that the older breakers leak a bit more than the newer breakers. But there's enough breakers of a wide enough age range that even a very low leakage rate is going to result in some breaker or another getting to low every few weeks.

 

[submonkey]

Hi Mbrooke,

We trip via the protection relay because:

- The densimeter contacts are often not rated for tripping.
- We want the relay to flag and collect an oscillographic record.

Breakers which trip on their own without flags would be quite confusing
for operating staff in my opinion.

Some manufacturers do offer an option to trip directly from the densimeter.

Breakers are not supposed to leak. In Australia utilities are required
to report on quantities of leaked SF6. Leaking breakers should be either
repaired or replaced.

The only credible mechanism I've found for a fast leak is maloperation
of a pressure relief disk, which occurs very, very rarely.

Thanks,
Alan

 

[Mbrooke]

@Submonkey: Thanks and very interesting. What advantage does an oscilliographic record offer? Is it to determine if the arcing was extinguished fast enough under low SF6? IF so it sounds like there is indeed a concern regarding tripping under low pressure rather than locking out. As for the density meter, would it not be possible to have a relay inside the breaker cabinet that closes forcing the trip solenoid to open along with closing another contact simultaneously to trigger a remote annunciator/indicator that the breaker has opened under low SF6? This way any confusion among personnel would be cleared up.


The breakers in Australia are live tank? Or do you have any dead tanks?

 

[Mbrooke]

We get a steady stream of low gas alarms, all of which get attended to post haste. To my knowledge we've never had a breaker fail to trip due to the low-low block.

How many SF6 breaker do you have in service territory? If the number is large, this sounds like an excellent reliability and track record.

An automatic trip at low would be highly disruptive, possibly dropping load.

While I am not familiar with your utility practices, in my application this is usually not a major concern. In the vast majority of cases all load serving substations are fed with at least 2 transmission sources, and the vast majority of step down transformers are operated in parallel with reserve capacity and their own circuit breaker, so a surprise trip does result in outages. At medium and large generation stations redundant breaker schemes are employed (double breaker double buss) so the trip of any breaker does not remove generation provided no other work is being done. But even then the system is always run such that the unexpected loss of any two generators within several minutes will not result in a major frequency dip.

The switching necessary to isolate a breaker to add SF6 can be done without dropping load. Any breaker that goes from low to low-low faster than we can get gas added to it is probably loosing gas too fast to safely trip below the point at which it is no longer safe to allow the breaker to trip.

Perhaps, but correct me if I am wrong. I am reading that even at near atmospheric SF6 levels a breaker can interrupt rated full load load. So provided the breaker is not opening a short circuit, no risk exists on tripping the breaker at low-low during a fast leak?

A clean breaker failure trip following a breaker at low-low having the trip blocked will be recovered from much quicker than the mess following a catastrophic breaker failure. Do what you will, but we'll continue to not trip on gas alarms.

My understanding from everything I gather thus far is that provided the breaker is not leaking as a result of vandalism, no risk exists from tripping on low-low?

 

[Mbrooke]

I just want to add this here for anyone reading up in this topic.

http://www.nerc.com/pa/rrm/ea/Lesso...r_Modification_Leads_to_Inadvertent_Trips.pdf

 

[submonkey]

Hi Mbrooke,

I try and ensure that an oscillographic record is triggered for any automatic
trip operation. Even very simple circuits can malfunction, and it is always
useful to know what was happening around the time of the event. For example,
the densimeter may get bumped or knocked, the breaker may fail to trip when
the trip coil is energised, etc. I am also interested in the primary current
waveforms associated with interruption - if a fast leak does occur (and it
might), it'll be useful to know how the breaker handled the situation.

Your proposed circuit where the tripping is initiated in the breaker cabinet
is fine - there are many ways to achieve the same end.

Australia has multiple utilities, probably all with different mixes of switchgear.
We have a mix of live tank outdoor breakers, dead tank outdoor breakers, and GIS.
Most of our outdoor breakers are dead tank, and dead tank breakers are used whenever
practical on new/replacement jobs. Live tanks are slowly going away.

It must be stated that tripping on stage 2 low gas *does* carry some risk.
If a fast leak occurs (e.g. a rupture disk spontaneously bursts), the breaker
may fail to interrupt, and such a failure could easily be catastrophic. Such
a failure would typically result in operation of the busbar protection.

It is not correct to assume that a breaker will interrupt load with the SF6
density below the stage 2 level. For some breaker designs, interrupting a
low current may even be more "difficult" than interrupting a high current.
Many breaker designs use the energy from the arc to assist with interruption.

The alternative (locking out) also carries risk. The densimeter and lockout
circuitry add failure modes to both protection circuits. Additionally, a locked
out breaker puts the network at risk until a human response occurs. Operational
staff may miss alarms, the alarm may not function as required, etc.

On balance, while I certainly worry about the "fast leak" possibility, I think that
the overall least risk approach is to trip. I don't think spontaneous operation of
a rupture disk is any more likely than a spontaneous failure of insulation within
the breaker, which is an accepted risk.

The impact on reliability is also a non-issue in my view. I see several breaker
operations each week due to genuine faults, and the network survives. There are
only a handful of stage 1 gas alarms each year, and only a tiny fraction of those
make it to stage 2 before human intervention (one every few years). Trips due to
low gas are therefore insignificant compared to "normal" network faults.

There is no perfect solution - you have to weigh up the risks and make a decision.
If you choose to lock out, I'd strongly suggest specifying two densimeters (one for
each protection). Your switchgear vendor will complain, but will comply if you insist.

It is interesting to me that protection engineers pay such close attention to the
circuity inside protection panels, yet often do not even look at the circuitry that
the switchgear people are putting in the breakers.

I also note that nobody locks out bulk oil circuit breakers when the oil level become
low. I've seen as many breakers leaking oil as I've seen leaking SF6. We're an odd bunch.

Thanks,
Alan

 

[stevenal]

Perhaps, but correct me if I am wrong. I am reading that even at near atmospheric SF6 levels a breaker can interrupt rated full load load. So provided the breaker is not opening a short circuit, no risk exists on tripping the breaker at low-low during a fast leak?

That may be true for pure SF6, but leaks work in two directions. At atmospheric pressure, you have a mixture of SF6, air, and water vapor. I would not be counting on it to interrupt.

In the vast majority of cases all load serving substations are fed with at least 2 transmission sources,..

Great policy, it makes maintenance so much easier. So how do we ensure the low low event doesn't occur during the time one source is out of service for maintenance?

 

[Mbrooke]

That may be true for pure SF6, but leaks work in two directions. At atmospheric pressure, you have a mixture of SF6, air, and water vapor. I would not be counting on it to interrupt.

Point taken

Great policy, it makes maintenance so much easier. So how do we ensure the low low event doesn't occur during the time one source is out of service for maintenance?

Outside of redundant breaker designs (BR 1/2 & DBR-DB) you bring up a good point. If the leak is slow and the breaker alarms first (on low) it is possible fill it or take it out of service without interrupting load (transfer buss). Of course this assumes a very slow leak like DavidBeach brought up along with a fast enough response to take care of it.

However, view it from this vantage point. If during maintenance an arc flash occurred on the neighboring circuit or bus bar at the receiving station, line protection is disabled and will require clearing the source (supply) substation.

 

[Mbrooke]

Choice has been made to trip on low-low.

But: once I trip on low-low, do I need to open the disconnects in between the breaker to remove voltage or can the breaker remain "energized"?

 

[submonkey]

Hi Mbrooke,

I'm interested in how you made your decision - was it your choice alone
or was there a reasonable amount of debate within your company?

I don't think opening the disconnects is essential (hand operated
disconnects obviously require human intervention). If you can easily
open them, it's nice to do.

Thanks,
Alan

 

[Mbrooke]

There was debate that went in to this, and Id also like to thank everyone who shared their experience on the matter.

Minus deliberate sabotage (which has never been documented around here and not foreseen) a fast leak is not a realistic concern holding good probability. A breaker tripping due to a malfunctioning density meter is considered a desirable failure mode over disabling all protection where in some stations this results in a considerable number of circuits being tripped via breaker failure. Also a bad ice storm or sever weather event where access can not be gained to a substation: a breaker that won't close is preferred over one that won't open. To prevent a breaker being energized and suddenly tripping on dangerously low SF6 operating procedure requires varying healthy 6FS pressure to any breaker about to be re-energized or re-closed and as a last resort once a breaker reaches low-low the close coil (and possibly motor)will be disabled.


The argument was made in what case would a breaker be allowed to go to low-low when it would receive immediate attention in the low stage, and while this was the strongest argument by far, one has to realize some smaller stations do not have adequate SCADA alarming (being old and built around pampering bulk oil breakers lol) and thus a low stage can only be noticed manually in between inspection.

Ideally- and this will probably happen in the future with "smart grid"- a rate of change meter giving rich pressure feedback to service members and SCADA control. A rapid decline would lock the breaker out, while a gradual leak will trip on low-low as desired. The issue is cost and how to go about this technical leap in the present time frame.

All the breakers in question are dead tank btw- 72.5kv to 362kv ratings.

As for disconnecting the breaker it is doable where MODs exist and done at half of all stations. There is a debate whether to keep on MODing or not, BUT, such will happily sway the argument toward MODing.

Where I am stuck is understanding what risk a breaker remaining energized holds.

 

[marks1080]

I'm used to the breaker tripping itself out for low-low SF6. What type of breaker are you using? I'm thinking all of your discussions have been moot as the breaker will not wait for a trip signal on low-low SF6. I suggest you really go through the breaker control circuit to see how it behaves on its own. It will trip itself. Above you mentioned that a breaker was 'allowed to go into low-low SF6 without tripping.' This is very confusing to me. Has someone tampered with the internal breaker control cct to allow this? If not, you may want to look at buying different breakers.

Alarm on low SF6. Depending on voltage level you may want to move disconnect switches to really isolate the breaker (assuming you have MO disconnects) on low-low SF6.

It should also lock itself out. Out of paranoia (or added security) you will often see these signals go into the relaying world. Any action to trip the breaker via protection will be redundant. Any 'block reclose' signal or supervision added into your reclosure scheme will be redundant. In my opinion you DO want to initiate breaker fail - though I've heard debates around this, I personally feel like breaker fail is required, especially when you already know that breaker is not healthy.

The discussion around being able to trip vs close a breaker should never occur at the same time. 'Tripping' is very different than 'opening' and totally different from 'closing.' Open and close commands to a breaker are operational procedures. Tripping is protection. Protection is what 99.9% of your attention should be focused on. A P&C engineer has to always make sure a breaker can trip during a fault.

 

[Mbrooke]

I'm used to the breaker tripping itself out for low-low SF6. What type of breaker are you using?

Dead tank, Siemens.

I'm thinking all of your discussions have been moot as the breaker will not wait for a trip signal on low-low SF6.

As soon as low-low SF6 is struck its trip solenoid with block close.

I suggest you really go through the breaker control circuit to see how it behaves on its own. It will trip itself. Above you mentioned that a breaker was 'allowed to go into low-low SF6 without tripping.' This is very confusing to me.

Sorry, my mistake. If for what ever reason both DC circuits to the breaker have been disconnected, and the density meter closed on low-low during that DC circuit off period, should the circuit be re-energized afterward it would result in immediate tripping. Of course, that trip could take place with zero SF6. For this reason ANY breaker being energized on the HV bushings or control circuit MUST be visually inspected to assure ample SF6 gas pressure. A breaker with low pressure will not even by considered safe to put back in service.

Is this what you had in mind?

Has someone tampered with the internal breaker control cct to allow this? If not, you may want to look at buying different breakers.

Not yet, but its being spoken about ordering new breakers from the factory wired for spontaneous trip and block close from the factory.

Alarm on low SF6. Depending on voltage level you may want to move disconnect switches to really isolate the breaker (assuming you have MO disconnects) on low-low SF6.

As long as the breaker stays open and can not be closed electrically in any shape of form, is isolation necessary?

It should also lock itself out.

OH YES! As soon as low-low initiates, the close coil (and spring charge motor if doable) will be disabled from electrical operation.

Out of paranoia (or added security) you will often see these signals go into the relaying world. Any action to trip the breaker via protection will be redundant. Any 'block reclose' signal or supervision added into your reclosure scheme will be redundant. In my opinion you DO want to initiate breaker fail - though I've heard debates around this, I personally feel like breaker fail is required, especially when you already know that breaker is not healthy.

Breaker fail for opening automatically? When the breaker trips on low-low an alert is sent via into the control hut and via SCADA. Like this field personnel will be aware that the breaker has tripped due to low SF6.

The discussion around being able to trip vs close a breaker should never occur at the same time. 'Tripping' is very different than 'opening' and totally different from 'closing.' Open and close commands to a breaker are operational procedures. Tripping is protection. Protection is what 99.9% of your attention should be focused on. A P&C engineer has to always make sure a breaker can trip during a fault.

Not sure where it stands, but the idea is to initiate the trip coil the millisecond that densimeter hits low-low, electrically isolate the close coils and alarm that the breaker has tripped on low-low. Nothing can close that breaker either relay or SCADA. At this point we are happy in theory.

 

[marks1080]

Hi Mbrooke,

I know that these breakers are generally rated to open with low or no SF6 for load currents. However, I personally think it would be a huge mistake to use this fact in operating decisions. If a breaker is known to have low-low SF6 it should be taken out of service and isolated immediately. Look at it from the point of view of how many possible contingencies away from catastrophic failure you are. If you don't isolate the breaker when you know its unhealthy how will it perform if exposed to some other system fault or condition? Treat the breaker as being 'failed' before it actually blows up. Initiate breaker fail on low-low gas and let the breaker fail protection clear the zone. Don't clear the zone if it's not necessary to do so. Do not modify the manufactures lock-out and trip logic. I'm fairly amazed those discussions are even taking place at your company.

Use the 'SF6 Low Gas Trip' contacts to initiate breaker fail. For these scenarios you need to rely on the breaker tripping itself. Sending signals via SCADA is an operational practice, nothing to do with protections. If you know your breaker is unhealthy (low-low SF6 for example) the fastest way to trip it (faster than BF or any other protection) is to let it trip itself. You are only a single contact operation away from picking up the trip coil then, so you trip speed is the speed of your motorized SF6 pressure switches plus one contact. If you trip via any other system you will add extra time to the trip action. You initiate breaker fail to take care of clearing the zone in the even that the breaker can't break the current, or for some other concurrent failure. Remember, when things go bad in primary equipment you can reasonably expect the chances for concurrent failures to go up.

For stage 1 low-SF6 you should alarm and dispatch. No need to take the breaker out of service immediately, although the dispatched workers who inspect the breaker would probably appreciate an outage. This can all be done via operator control, and no protection operation necessary.

One last point on Breaker Fail: I'm not sure but I got the impression from your above response that BF opens said breaker. This is not true. Breaker Fail does not try to trip it's own breaker, only the breakers in the adjacent zone. There's actually a really good reason not to have a breaker fail protection send a trip into it's own breaker, and it has to do with making sure you don't leave your 125/250V DC signal on the trip coil, eventually cooking it to destruction.

 

[FrozenE-]

I know that these breakers are generally rated to open with low or no SF6 for load currents.

Can you provide some further information on this one? It's something I've never come across before and I'd be surprised if any breaker manufacturer would stand behind that statement.

There's actually a really good reason not to have a breaker fail protection send a trip into it's own breaker, and it has to do with making sure you don't leave your 125/250V DC signal on the trip coil, eventually cooking it to destruction.

The only time I've seen coils burn up is because the breaker failed to operate at all (mech problem, coil fell off the solenoid, etc). If the trip is being blocked by low gas, low spring, air, etc, the coil won't burn up - current will never get to it. Hence locked-out.

And for my own clarification, this:

signal is send to breaker failure protection for zone fast tripping in case of fault on this feeder.

Is not the same as this:

In my opinion you DO want to initiate breaker fail - though I've heard debates around this, I personally feel like breaker fail is required, especially when you already know that breaker is not healthy.

Are you suggesting that on top of having the breaker trip due to low gas, that the breaker fail protection clear the affected zone? That could get ugly pretty quickly.


If we're tallying, my utility does NOT trip on low-low.

 

[marks1080]

FrozenE,

I'm suggesting that you initiate breaker fail every time you trip the breaker - whether it's an automatic trip by the breaker, or a protection trip. Remember, tripping and opening are very different things. Simply initiating breaker fail doesn't clear the zone. That will only happen if breaker fail conditions are TRUE. Then wouldn't you be glad that you initiated it?

You can check with different breaker manufacturers directly as to weather or not they will stand by their breaker opening during a low-low gas scenario. I can't say that all manufacturers provide for this, although it would be very surprising if they don't. This all part of the BIL rating of the breaker. I can tell you that the manufacturer of an ABB 145PM63-20B has a specified 750kV open contact LIL. As per the spec sheet. As a P&C guy, I would also suggest talking with your in-house equipment engineers as they are generally far more knowledgeable about all the very detailed specifications of the breaker.

I can say that all the dead-tank SF6 breakers I've worked with (mitsubishi, siemens, abb) at the 115kV or higher level all trip themselves out on low-low gas. Avoiding this would require modifying the manufacturers wiring from what I've seen. It's kind like messing with an old oil breaker anti-pump or anti-slam scheme. You just wouldn't do it. Well not if you really understand the consequences of doing so.

 

[FrozenE-]

Ok, my miss-reading of your comment on breaker fail "initiation". Fair enough.

As for the tripping at low-low - I would not consider low-low to be near a point where there is "low or no SF6". Especially the NO SF6 part. I agree, at low-low they will still trip adequately; if they didn't then the low-low point would be too, well, low. There were some comments above about a breaker being operated at near atmospheric pressure -

Perhaps, but correct me if I am wrong. I am reading that even at near atmospheric SF6 levels a breaker can interrupt rated full load load.

which I would definitely not agree with.

I've commissioned a half-dozen 145PM's (as well as a good deal of other breakers) and none of them had the wiring modified by us to prevent a low-low trip - they all came as such from factory. The "general" schematics (including the ones that say to make sure you refer to the schematic furnished with that specific breaker to ensure it is correct) all show a lockout at low-low, not a trip. Our experiences are not the same, it would seem - who would have thought?

Thanks.

 

[marks1080]

FrozenE:

Ya it's surprising how different procedures are from region to region, or even company to company.

I've always thought that the breaker should be able to break for BIL rating at atmospheric pressure, but I concede I could very well be wrong on this. I've never gone further into it, because I don't think it would affect my decision making. I generally take the overly cautious approach when dealing with protection equipment. I try to design for different contingencies, assuming there's a 2nd contingency scenario also happening. For example... If I'm designing how to handle low-low gas signals from a breaker (1st contingency) I will use that information in my protections such that I'm sure that a concurrent contingency (line fault for example) would not compromise protections. By that I mean, I want to minimize the amount of breakers tripping for each contingency scenario.

Its common for different customers to get different breaker controls ccts coming with their breakers depending on what each customer has already asked customization for. Whether or not my company asked for the low-low trip or your company asked for it to be removed, I wouldn't know. I would assume the later though. Manufacturers are interested in making sure the equipment they sell doesn't blow up. They would be motivated to provide a self trip on low-low gas. The 145PM63-20B control cct i have from the manufacturer in front of me right now clearly shows a self trip on low-low gas driven by the 63-2 pressure switches, independently in BOTH trip coil ccts.

I believe that companies who are more focused on operational engineering may push to remove the self trip... just a guess though. I would personally not put a breaker in service without it. Have you worked with air-blast breaker that do the exact opposite action on low-low gas? Again, I would be surprised to see an air blast breaker that didn't close itself on low-low gas, just like I would be surprised to see an SF6 not trip itself on low-low.

Great discussion either way. It's interesting to see how different companies use the same equipment in different ways for different reasons.

 

[FrozenE-]

Absolutely, and exactly why I've been lurking around this site for years.

My 63-2's show the trip cct being blocked open (both cct's of course). Interesting indeed - I agree, our companies could have spec'd as such, I'm in support, not supply - but I will be enquiring when I get a chance.

I have indeed been involved in the nightmare that is air blast breakers. Ours do not automatically close on low air, but they sure do it by themselves once the air pressure gets low enough. It's a major concern here in Canada, as seals don't work to well at -30 or -40C. If I understand you correctly your ABCB's intentionally initiate a close while there's still enough air to operate? Very interesting indeed.