Arc Flash Detection 1

[X49]

I'm considering installing arc flash detection equipment to trip upstream breakers instantaneously in the case of an arc flash event. My question is whether this equipment can be of any use in instances where the arcing fault current is already expected to be high enough to trip the breaker on instantaneous.

For example, I've got a 1000A breaker which will trip on intantaneous at 10kA or less, and the calculated arcing fault current is 15kA. Will the arc flash detection trip the breaker sooner due to the time it takes for the fault current to build up? Or will the system be completely useless for all types of faults in this case?

 

[ScottyUK]

The arcing fault current is 15kA? Or the bolted fault current is 15kA? Arc current won't often reach the bolted fault level.

If your arc was 'only' 7kA or 8kA would you want to initiate a breaker trip immediately or wait for the thermal trip to clear it?

 

[X49]

The arcing fault current is expected to be in the 15kA to 20kA range based on IEEE 1584 calculations. Bolted three phase fault current will be higher than this.

However, I do not know if there could be situations where the arcing current could be lower than the theoretical calculated value. Also, I do not know if it could take a significant amount of time for the arcing current to build up to 15kA.

Also, if any mods are reading, this thread should probably be moved to Electric Power & T&D.

 

[jraef]

If you were thinking to use the light operated detection systems, the entire purpose is to act FASTER than any current detection can, so as to initiate the clearing process fast enough to limit the fault current. But without you stating what you were planning on using, it's difficult to answer in any greater detail.



"Will work for (the memory of) salami"

 

[davidbeach]

In theory you could get the trip signal to the breaker up to about 12-20ms faster than an external relay instantaneous element could, but probably slower than an internal magnetic instantaneous element. Actual results could vary.

 

[X49]

jraef, The system is a light operated detection system.

david, From what you're saying, there would be not be much of an advantage to having the detection system, since the breakers will have internal magnetic trips which should be as fast as the optical relays.

 

[davidbeach]

Yep, the light systems only really provide a noticeable decrease in tripping time if there is no other instantaneous element that's going to respond to the fault. Switchgear without bus differential is one example where the light sensing can make a real difference. On the other hand, the light system won't respond to a bolted bus fault while bus differential would, so I'd just go with bus diff and leave out the optics.

 

[bacon4life]

The SEL-751 specification lists a pickup time of 2-5 ms for arc-flash overcurrent & overlight elements but a slightly longer pickup time of <1.5 cycles for instantaneous overcurrent elements.

 

[dpc]

If you have instantaneous fault clearing or differential protection, I don't think there is much justification for adding the arc sensing. I also have concerns about reliability of these systems in detecting every possible arcing fault that might occur inside the gear. Even if the arc sensing is slightly faster, the breaker clearing time of 5 or 3 cycles must be added to the relay operate times in either case (plus 86 operate time if applicable), so this somewhat levels the playing field.

 

[living2learn]

When forced to use this stuff, I like to wire up the door contacts to disengage the light arc flash detection to prevent photos, outside light from tripping/locking out your switchgear. I agree that a MV breaker takes 3 cycles for the new ones, so

 

[dpc]

I don't see too much point in having the protection if you are going to disable it when the door is open. The arc detection element needs to be supervised by an overcurrent function to reduce the risk of nuisance tripping. This is provided for by both the ABB and SEL relays that I am familiar with.

 

[living2learn]

was using the arc flash as more of a bus protection vs. personal protection

thoughts?

 

[davidbeach]

Bus differential does a far better job of bus protection than a light based arc flash relay will. Lots of things that can happen that are very bad for the bus and will create lots of differential current but no light.

 

[Gragas]

Light detection relays are typically fairly cheap and are indeed faster, thus typically reducing arc flash levels.

Arc Maint switches are also useful and extremely cheap.

Bus differential is fast, but I disagree it does a "far better job". Its also typically more expensive. Bus diff used to be the end all be all. It all depends on the Plant preference. Also, there is no reason to assume everything is the sole item protecting the bus.

Its all just a box of tools, choose based on Plant, or more importantly in the arc cases anymore, or insurance requirements/preferences.

There are a couple of white papers out that compare times of different protection elements/philosophies. It has been a few days since I found them but a decent google search should turn them up.

 

[davidbeach]

Ok, then how about "Bus differential does a far more complete job of bus protection..." The light requires the fault to be arcing in a compartment where there are light sensors, bus differential merely requires the current into the zone to resemble a bus fault; no arcing required. If 12-20ms faster clearing makes a significant difference in the outcome of the arc-flash hazard calculations, you're putting too much faith into the results of calculations that have lots of empirically derived factors. I wouldn't really trust my life to a calculation that lets me into a lower level of PPE by 0.05 cal/cm[sup]2[/sup] based on a 20ms faster trip.

 

[Gragas]

Its larger drop than you would think.

I have never seen a light detection scheme were the detection isn't easily routed through the compartments.

Also, I do not diagree there is a bunch of empirically derived factors. However, the "reliance" is typically strictly dictated by Plant safety requirements, insurance requirements and contractrual details.

These typically specify levels and values specifically, making the "reliance" fairly strong.

You only have so many tools at your disposal. As for "complete" protection... again, I would not assume that its strictly one vs. the other. Typically there is more than just one or the other on the bus.

 

[squeeky]

Balance your risk against the cost. Are you protecting people? Are you retro fitting to old gear? I have seen many types but have only installed tha ABB one. They offer different modules for different applications. The master units "talk" to each other via fiber links. Lower level units can be used to "look" at cable compartments. Configuring them together can allow zone tripping to limit fault areas. Thank God I have never had a fault where it has worked but have seen compariable faults with and without. I'd rather have it. I would retro fit to old swich gear if there was no money to upgrade to new.