This is my first post, and please excuse me if this question has been asked. I haven't figured out how to use the search function.

I am doing a series of arc flash studies using ETAP 18.0. My client has asked that I include all busses above 50V, which would include all lighting panels and 120VAC UPS loads. I informed him that IEEE 1584 states that it's not necessary to include items below 240V unless they are on a transformer of 125 kVA or greater. He told me that IEEE 1584 was going to change this. I don't know if this is true or not. I am not aware of this.

Are there any accepted calculations for doing arc flash studies on 120VAC systems with very small transformers? I'm referring to a single phase 240-120V lighting panel on a 10 kVA transformer. This is less power than at a typical house. In addition we have numerous UPS systems, all 2kVA. Studies are also expected on these systems.

Before I dig into this I thought I'd ask if anyone is aware of industry-accepted calculations for these two situations. And if so, can someone please send me a link for these?

Thanks
Andy

Hello,
First off, Welcome to the Forum!

At this point it is unknown to the general population on what is in the new IEEE 1584 proposed draft. There have been some discussion on it in this forum.

The only choice you have is to either use the exception or use the 3 phase calculation for the single phase value. This will result in conservatively high results. I think you will find that the calculated incident energy will be very high for a single phase system. You might want to run a couple of those and show the results to your client and ask if that would acceptable to the electricians to wear 20 cal clothing for a 120V system.

i believe everyone uses the exception for single phase 120V systems.

Basically the problem with IEEE 1584-2002 edition is that there is a single test result at 208 V. ALL other tests that were attempted at 208 V failed to arc long enough to be considered valid data, and that's the big problem. IEEE 1584 empirical model is based on stable arcs, nothing less. Playing around below 250 VAC or so runs afoul of this stability problem.

IEEE 1584 Draft 6 is out there on IEEE Xplore. This is a DRAFT so of course it is still subject to change. The language in Draft 6 is more like <250 VAC with <2 kA of available fault current. You can read it off that web site if you pay for access. Running some basic infinite bus calculations, your examples fall within this window. BUT this is still a draft. You really can't be staking your study based on a draft document. Draft 6 does also extend the "valid" range down quite far but they did some tricks in it like ignoring the first cycle before taking data so that the model applies to arcs during their stable condition. However it still does not quantify when arcs are stable and when they aren't.

So coming back up for air, there are basically two ways to approach it to date that are based on standards. You can either use the "clause" in IEEE 1584-2002 (<125 kVA, <240 VAC) or you can use the table based method in IEEE C2 which is based on actual equipment testing. In that method since it is for utilities they start with an assumed arc flash PPE of ATPV 4 across the board with no face shield as a bare minimum PPE. Any equipment rated 250 VAC or less qualifies for this level of PPE regardless of any performance criteria. I believe the highest level of incident energy they achieved in laboratory testing was 3.2 cal/cm2. Second option as stated earlier is just use IEEE 1584 at least down to 208 VAC as three phase. Can't help you if you are outside of that range. Doan or Lee models are your only choice. Lee in particular is actually an IEEE 1584-2002 "model". It's just theoretical instead of empirically based.

OR and this is my best suggestion, either use the paper I put on this web site in the articles section on 250 VAC or less arcs, OR virtually the same paper that the folks at ETAP pulled together as a white paper although it was written independently of mine. They also included a lot more modelling (in ETAP of course) to prove the point. For example one of the results you can use is that Duke commissioned tests in support of 125 VDC substation batteries when they tested 130 VDC arcs. At 20 kA DC or less of available fault current at 130 VDC it just kisses the 1.2 cal/cm2 threshold. Thus we can be absolutely assured that AC arcs at these voltages simply won't be nearly that high since AC arcs ignite and extinguish 120 times per second compared to DC arcs that do not, using RMS values of course. This pretty much eliminates any and/or all 110-120 VAC systems at 20 kA or less from consideration.

Right at around 200-250 VAC is the threshold for stable arcs. This is where it gets tricky and why it gets tricky. There is no decent comprehensive study of arc stability to date. The joint IEEE/NFPA study did some preliminary testing that mostly just validated what was done in the past then moved on to stable arcs without returning to the conditions in question.

Thank you both. I have run the ETAP model a few times, using their single phase arc flash function, and the results are not believable. In my model I have a 750 kVA transformer feeding an MCC. The MCC has buckets for several motors, including 2 @ 250 Hp. Also included is one bucket with a 40A breaker for a 15 kVA single phase lighting transformer, 240/120VAC. With the exception of the incoming MCC main, the lighting panel downstream of the 15 kVA transformer has a higher IE than the entire MCC. The model shows that the 400A breakers for the 250 Hp motors are safer than a 15A, 120VAC lighting circuit. Intuitively I can see that this is wrong, but I'd like to know why it's wrong before I ask my client to reconsider this request. I will look on the site for this paper being discussed. That will be a good start.

Thanks again
Andy

Actually the result is probably right, it just may not be very intuitive.

First let's suppose for a minute that we're looking at this purely as a three phase system. Now I don't know if your result is based on modelling the secondary or primary side of the transformer but let's work with the primary side first. We need to know the arcing current and compare it to the curves for the 40 and 400 A breakers. It depends a little on where on the TCC that the arcing fault current falls. Simple as that, so most likely this is due to miscoordination.

Now let's consider the secondary side of the transformer. No secondary protection mentioned so I'm assuming that this is on the secondary leads of the transformer. Since the transformer is an impedance, this lowers the arcing current on the secondary side. Intuitively it would seem that if arcing power is reduced, then arc energy and thus incident energy would also be reduced.

This is certainly true if the arcing current is within the breaker fault curves. But there's that pesky time element in there and what causes a problem. If we are in the instantaneous time region of the circuit breakers with the primary and secondary cases, then as expected arc power goes down while the opening time remains fixed. That's the easy one but clearly this just disproves the calculations that you've mentioned. HOWEVER if we are within an inverse time curve region, things don't work that way. The linear reduction in arc power also corresponds to an exponential increase in arcing time. The net result is that incident energy thus increases rather than decreases as the transformer impedance increases. This is VERY commonly the case with circuit breakers and fuses for transformers because the breaker/fuse size has to be pushed up very high to overcome effects of inrush when the breaker or disconnect is engaged.

Now the monkey in the wrench here is that ETAP is using some different equation for 1 phase circuits, and uses different data. This isn't an IEEE 1584 standard so you'd have to look closely and validate it yourself.

From a mathematical stand point, I agree with you. From an engineering standpoint it is clear that this is wrong. The basis of this analysis is dependent on the arc fault being self sustaining, which is very unlikely. The equation calculates energy, even when it's not there.

Imagine doing the same analysis on a 1 kva transformer with a 120V secondary. If you protect the transformer with a large enough fuse, a fault on the secondary will never clear. That doesn't make a fault on the secondary of this tiny transformer an arc hazard, even though the analysis would indicate it is. It's a misapplication of the equation.

I agree 100% with this statement. If we assume stability, we get the ETAP result. Mathematically it's correct though it might be a little confusing because the result is based on energy which is power multiplied by TIME and in this case both power and time are changing in opposite directions. The good news is that we are pretty sure intuitively that the arc is NOT stable so this is really just an upper bound. The real arc flash hazard won't exceed this number, but it leaves a bad taste in the mouth, a lot more inspection, and it's obviously just the wrong approach.

But if we assume or try to estimate stability, we're in trouble for the simple reason that we don't honestly have a comprehensive standard. We have very few options to go by. Let's take them in turn. First, we have IEEE C2 (NESC). This is a utility standard and the front matter clearly documents that it is not really intended to apply to industrial settings as a general standard so we have to overcome and accept this. But it has a very simple table based approach that gets us to ATPV 4 without face protection except a hard hat. This is perfect if you are working in a metal casting shop, a refinery, natural gas plant, or similar operations where NFPA 2112 for flash fires rules and you are already wearing that level of PPE for other reasons but it is problematic for all other industries where FR PPE is not the norm for daily work wear. Some plants have simply accepted this and the electricians are required to wear it though so it's not an impractical standard. Personally I go pretty much everywhere. My uniforms include FR pants and I have a choice of a short sleeve non-FR shirt or a long sleeve FR shirt, and I carry a 40 cal suit on the truck in case that's not enough. I also carry a full change of clothes and I even have an old Hi Viz+FR lineman's shirt for those very rare situations where I need both. I think I've got both FR and regular hair nets too. It also comes in handy if I end up having to stay overnight or I get so dirty I need to change to drive home. This covers me with all my clients. All line crews around here work every day in the hot summer sun in the South in full 4+ ATPV uniforms. They hated it when it changed but they've learned to live with it. There is also a plant in Trinidad that I know the engineer at that has the same standard as every day workwear for all of their electricians and they've had no increases in heat related illnesses. The advantage of this approach is not only is IEEE C2 revised a lot more often (5 year cycle) but the results haven't been thrown into question by test results. The downside of course is there is always minimal PPE.

But if we want to get to NO PPE, our options are a little more limited. Right now there is a consensus safety standard, IEEE 1584-2002, that gives a specific criteria for no PPE which is the "<240 V, <=125 kVA" rule. Just as with other consensus standards such as NEC (NFPA 70) or NFPA 70E the normal result is that consensus safety standards are standards. They are however subject to review and change over time. On occasion an experimental result or some other new information will get published. This becomes public input which is used by the standards committee to be considered for review and revision of the standard that the committee maintains. The revised standard eventually gets published. Then the various organizations that base their results on the standard review and accept, reject, or modify, the new safety standard. Due to the extensive review process itself, consensus safety standards have the legal weight of just below regulations. So you could ignore all the machinations and simply hold fast to PUBLISHED, APPROVED, FINAL consensus safety standards. That's the normal way things are done.

Now as engineers of course we should review and accept, reject, or modify the underlying standards that we use as the basis for a decision. As an example if I'm working in the mining business in the U.S., the federal mine safety administration (MSHA) recognizes a part of the 1976 edition of NEC as part of their regulations. Mines fall generally outside state regulatory authority. This means that unlike most commercial/industrial jurisdictions, mines are really only required to meet the 1976 NEC standard. They can freely adopt or ignore NEC. Most more progressively oriented mines adopt a more current NEC standard as an engineering and safety standard though because they recognize that the 1976 standard and the mine regulations leave a LOT to be desired. This would be equivalent to following OSHA 1910 Subchapter S for general industry...it is certainly a regulatory requirement but far from a best practice today.

The danger of course is that if you develop your own standard, you are on your own to defend the underyling basis for that decision as well as whether or not you followed it. This is a far cry different from following a consensus safety standard in which case from a legal point of view the consensus safety standard has more legal weight than a bunch of engineers arguing over test results that were done in a lab and not necessarily verified or subject to an extensive review process. However there is a way to marry the two and keep yourself out of trouble. If you modify the standard to use a set of requirements that is more strict than the underlying standard, and you state that is in your standards review, you can pretty much freely implement any modified version of the underlying standard you want. This allows you to use experimental results directly, or use interim drafts of IEEE 1584, or simply pull something out of thin air because at the end of the day from a legal point of view, you are following IEEE 1584-2002 OR BETTER so IEEE 1584-2002 is the fall back support for this approach. As long as the path to proving the "OR BETTER" part of your criteria is easily followed, you honestly can't go wrong. So if you want to take this approach IEEE 1584 Second edition Draft 6 uses <240 VAC, <2 kA available fault current as the cutoff but for obvious reasons don't take my word for it. Download and read if for yourself.

So again you have 4 options:
1. Ignore arc extinguishment. Use IEEE 1584-2002 (empirical or Lee as appropriate) for single phase using the guidance from IEEE 1584-2002 to treat single phase as 3 phase. I'd be a little more nervous about using ETAP's single phase model without some kind of documentation that traces back to some kind of standard since it may not carry the same weight as a consensus safety standard.
2. Follow IEEE C2 (NESC) which is probably the strongest support among standards based approaches right now although it has a minimum PPE requirement. Whole industries and many plants have done this.
3. Follow IEEE 1584-2002 as it stands. Let the review process run it's course like it normally does. The equations themselves aren't all that great and a big improvement is coming on those too along with enclosure sizes. It's been over a decade since the last revision in an active research area so no wonder the standard shows it's age.
4. Modify IEEE 1584-2002 based on a more restrictive approach based on something you develop yourself. The engineering basis for your standard is not really in question since you are essentially following IEEE 1584-2002. You can almost say anything you want here.
5. Strike out on your own? This seems like serious trouble. Granted since IEEE 1584-2002 empirical results among others is just that...empirical, and not based on theory, the only arguments other than consensus review are goodness of fit to the data. So it is possible to use say the simplified arc flash equations published on an IEEE forum by a respected researcher, but we're treading on thin ice here.

Paul, nicely stated! I concur and we follow the NESC method simply because I do not believe "no label" is really not an option.

This is getting a bit off this topic but I disagree with the "everything is labeled" approach. It's just too impractical. Even the NEC rule on what gets labeled is very generous and leaves a lot of equipment without labels. Where do you stop? Control panels? Lighting panels? Light switches? Light fixtures? It gets out of hand very quickly especially when unqualified people start demanding labels everywhere, which turns into a giant maintenance headache. The only tricky part is that unlabeled equipment that was never analyzed should be covered under the NESC or NFPA 70E table-based approach or doing the analysis on the spot, OR you make the practical assumption that everything unlabeled is below your plant's minimum PPE requirement (1.2, 2, 4, or 8 cal/cm2 depending on PPE rules and industry standards). The latter is the usual approach but then you have to be extremely vigilant that you've covered all equipment thoroughly.