Hi all,

I'm wondering how you determine if the main breaker compartment is "sufficiently" isolated or if the barrier is "complete" such that it is permitted to be used as the fault clearing device for the assembly? Specifically, I'm wondering about switchgear assemblies.

From reading the older posts, it seems that everyone agrees that the compartmentalization in metal-clad switchgear provides sufficient isolation for the main protective device.

Is it reasonable to assume that all metal-clad switchgear constructed to C37 standard will provide sufficient isolation for the main protective device?

What about Arc Resistant enclosures Type 2C (the 'C' suffix denotes arc resistance between adjacent sections)? Would requiring Type 2C in order to consider an isolated main protective device be an overkill?

Thanks, looking forward to the responses.

This is more anecdotal but...

I have not seen any evidence of arc flash propagating from one enclosure to another. The arc resistant gear cases are very extreme...they continuously arc without interrupting for a long duration and so the special ratings are for burn-through which only happens in my experience if you have a total lack of protection.

I have definitely seen propagation in the vertical buses of MCC's in two different arc flash investigations and this is pretty well documented. I haven't seen it but I've heard claims of the same kind of thing in panelboards and it stands to reason that the construction is similar so I'd expect the same kind of things.

I have not seen propagation from one section of an MCC to another nor have I seen it within switchgear, EVEN in metal enclosed gear that typically has little more than sheet metal dividers that are not particularly "sealed" from one compartment to another. In the most recent case and mind you this wasn't an arc flash, this was simply an RC surge filter where the capacitor bushing blew out and the end the oil mist ignited, there was one very large enclosure between the compartment where the capacitor was along with the tap changer, and a second compartment with a large dry autotransformer. The autotransformer was untouched and even the tap changer itself was mostly intact which was about 3 feet away but all the wiring in the area was scorched and the sheet metal on both sides was buckled out or blown completely off, even though it was quite well vented. In another case with a 5 kV Class E2 starter running a 2500 HP motor, it arced over in the disconnect switch which in this equipment was located at what I'd call about midway in the gear (very old Model 2 or 3 Square D gear). It vaporized one phase and melted another one pretty badly and a lot of the insulation and glastic was severely scorched and smoke damaged. There was a 4 foot long scorch mark ont he floor at the bottom in front of the vent. But the doors never blew off and they weren't really even bowed out very badly. Everything was entirely contained within this starter and didn't propagate at all through any of the conduits or other openings except for the vents I mentioned.

So no I haven't seen much evidence of needing "bomb proof" switchgear. I really haven't even seen any need for arc resistant gear since all that it does for the customer is to contain an arcing fault ONLY if it occurs during normal operation, which is not the time when most of them happen. Most arc flashes occur during maintenance activities and those happen with the doors opened so that arc resistance is nonexistent. The biggest advantage of arc resistant gear is that in a market that is basically very low margin, it boosts the margins by a considerable amount for the vendors.

Thanks for your reply! I have also noted that IEEE 1584 doesn't specifically call out switchgear as an example. For example, IEEE 1584.1-2013 Section 8:



This implies that IEEE 1584 considers a switchgear main protective device to be typically sufficiently isolated, but they would not say it explicitly for legal reasons.

Would you say that even metal-enclosed switchgear with large openings for bus connections/cable terminations would provide enough segregation to consider the main protective device isolated?

Cheers.

Keep in mind that my experience is with mining and chemical plant equipment. At least with the mining equipment I never saw much in the way of propagation happening. But that's welded steel construction in a metal-enclosed style system like you are talking about. Not much of any sort of "dividers" between sections and typically due to the size of the doors, it blew them off pretty quickly so that it kind of "vents" itself and that kind of ends things. It just doesn't get nearly as exciting or as much damage as you see in low voltage gear. I think part of it too is that you need all that extra clearance and everything is sitting on insulators where it is not sitting an inch or two off the enclosure wall so you can't get the same amount of localized heating/melting that you'd get with the other stuff.

Lets say there is arc resistant switchgear but here is a shipping split and between the split is nothing more than sheet metal spanning a gap of around 12". Otherwise it is just like two sections of switchgear very close to each other.

Would stand near this gap during a fault? or while racking a breaker in? The blast pressure could be huge.

I forgot to mention this is a real piece of gear rated at 34kV running 24kV.

I agree with Paul that horizontal propagation, through a barrier, is unlikely. If it is a factory design barrier you are probably OK. I have seen 480V switchgear with no barrier.

Hugh Hoagland's group has had a major breakthrough when it comes to arc blast: they've actually measured it in a very reliable way. CIGRE has also done some monumental work in modeling it. The models are specifically intended to speed up the design/engineering/test aspect of arc resistant gear but nonetheless the information is quite useful if you sort of invert it and look at it from the point of view of non-arc resistant gear.

Hugh's group pretty much documented one crucial aspect of arc blast. There has been basically two prevailing theories as to what causes it. The first one is simply that the air inside the enclosure is heated and following basic understanding of gasses with no vent, it pressurizes. The second theory is that arc blast is caused by expansion of copper as it changes from solid to vapor form. Hugh's group in the most recent IAS-ESW documented that the copper vapor theory is either wrong or is so minor that it plays no role. This also matches the CIGRE work which uses the same basis for their modeling.

Where does this leave us? Ignoring arc resistant gear for a moment, we start with an enclosure. The enclosure may or may not be "sealed" sufficiently so that it builds up pressure. All of the documentation that I've been able to find on arc blasts measure a maximum of around 8-10 PSI within the enclosure and according to the CIGRE documentation this also happens very quickly...within one cycle. The reason that the pressure limits at 8-10 PSI is that smaller enclosures rupture at this pressure. Larger ones rupture at lower pressures. This stands to reason...10 PSI acting on a typical 1U bucket door which would be roughly 20"x20" has about 4,000 lbs. of pressure on it. That's guaranteed to blow off the hinges. I have been unsuccessful at finding anything credible that is higher than that.

Outside the enclosure though there are no claims about pressure at all. In fact except for some vague acoustical measurements and the data used by Lee to estimate arc blast which in itself isn't really representative because it involved an arc across two electrodes facing each other peaks out at around 2 PSI for high fault current cases.

So where does this leave us? 1-2 PSI AT MOST is the highest exposure we can expect. Maybe in some strange circumstance where someone is sealed in (and I mean SEALED) while an arc occurs we could get up to 8-10 PSI but that's it. At 1 PSI ear drums rupture and there are plenty of medical cases of this occurring. At around 20-35 PSI depending on which report you read is the threshold where a fatality is possible and somewhere in the neighborhood of 50-100 PSI fatalities are almost certain. That comes from military studies on the effects of a concussive blast which is what we are discussing here. So ear drums are definitely a possibility. Getting knocked around is also not out of the realm of reason. But fatalities or major injuries are flat out not going to happen. Effectively the claimed fatal consequences of an arc blast are a total myth. I can back it up another way, too. Doan, Hoagland, and Neal documented roughly some 50+ cases of arc flash and also listed secondary effects. Loss of hearing was one of them but nowhere in there are concussive type injuries listed. And I've had private exchanges with others that actually get to go out and do accident investigations on a regular basis who have also said that they have never seen any evidence of concussive "blast" type injuries outside of for instance blown ear drums.

As to people being "thrown back"...I think that's a myth, too. Look carefully at any of the videos on the internet of actual arc flash events and carefully single frame through them. Arc blast as previously noted happens in about 1 cycle so if that's the case we should see someone standing still in one frame and then suddenly thrown a significant distance in the very next frame and then stop moving by the third frame, assuming that the video frames are 60 per second (1 cycle) or slower. However what actually occurs is that about 0.1-0.4 seconds later, the person is suddenly moved away from the arc flash, way after the arc blast is already done. That's not arc blast...that's the "freeze or flight" reaction kicking in and the person is "propelled" away from an obvious danger under the own power. It might seem like they were "thrown" to people simply because the freeze/flight function in the brain is processed way down in the cerebellum and brain stem...it's more of an instinctive reaction than anything. By the time that the cerebrum figures out what is happening and begins to assess the situation, the cerebellum has already moved the person out of danger.

Anyways this is all opinion based on analyzing the available data six ways from Sunday. Based on this analysis I see no reason to make statements about not standing somewhere out of fear of an arc blast with the exception that for instance if you are going to be working off a scaffold or out of a bucket or somewhere else that you might get knocked out of (or jump out of), it's best to be in fall protection/prevention equipment, that you should stand to one side to avoid doors that might get blown off when an arcing fault may occur, and that ear plugs are probably a good idea.