Arc Flash at end loads 1

[JensenDrive]

NEC 110.16 says arc flash labeling is done at "switchboards, panelboards, industrial control panels, meter socket enclosures, and motor control centers." I am not sure how to interpret "industrial control panels." Note 110.16 does not metion arc flash labeling the end load, but broadly speaking, one could interpret any box that an elecrician could open up and work on live could have an arc flash hazard label as well.

Of those actually doing arc flash studies, how are you interpreting "industria control panels" and how far toward the end load are you commonly analyzing arc flash?

 

[dpc]

If it has 480 V, and would ever be opened while energized, it has be evaluated.

If it is only 120 V and below, I recommend a "generic" arc-flash label.

 

[alehman]

"If it has 480 V, and would ever be opened while energized, it has be evaluated."

Just to clarify - the above statement is true for compliance with NFPA 70E. Keep in mind that 70E has not been adopted by OSHA and is not widely adopted as a local code. The NEC requires only generic labeling. But...

OSHA has stated the following:
“Industry consensus standards, such as NFPA 70E, can be used by employers as guides to making the assessments and equipment selections required by the standard. Similarly, in OSHA enforcement actions, they (70E) can be used as evidence of whether the employer acted reasonably."

See
thread238-188899
thread238-171682

 

[JensenDrive]

That thread "Arc Flash - Where do you stop" is one I orginated. This question was a twist on that one, trying to find out if people are using concept that says, "I (or my consultant) will do the 70E and 1584 analysis, but I / we will still stop at the MCCs because that is all that is all that mentioned in the NEC." I have a suspicion that is the case. I cannot say it is right, but I think it is going on.

I started this project with a pre-existing SKM plant model that encompased about 150-200 buses that stopped at each major MCC and panel. The plant peak load is about 3 MVA, and I think there are lots of facilities of that size in the country. The arc flash model project meant adding all the loads and panels hanging off of the MCCs etc. After adding all this equipment, the number of buses will be about 1300 ( I am still adding data) (and this is in spite of stopping at the secondary of each 120/208 xfmr < 75kVA). I suspect this is typical. In conversations with SKM I gathered that owndership of large bus SKM licenses that can handle a plant of this size, is relatively uncommon. I was wondering why, which led to this posting.

 

[Zogzog]

"Keep in mind that 70E has not been adopted by OSHA and is not widely adopted as a local code"

Until 8/13/07 anyhow, thats when this statement will be false.

 

[dpc]

I always try to imagine myself in the witness chair explaining things to a predatory plaintiff's attorney:

"So you stopped your analysis at the MCC for what reason exactly?"

It would be tough to defend that approach in the case of downstream VFDs, package control panels, etc.

I don't think you need to model every load downstream of an MCC - only panels and other equipment that could be examined or opened while energized.

You also have to realize that downstream of a molded case breaker or fuse, there is a limit to how high the arc-flash energy can possibly be, provided you have enough fault current. This opens up the possibility of determining a maximum arc-flash level without modeling every downstream device.

As far as the NFPA 70E not being adopted by OSHA, that is not a line of defense that I would put much hope in. OSHA feels that their "general duty" clause allows them to pursue cases where PPE is not properly provided and used and arc-flash injury results. They have already done so, and fines have been paid. NFPA 70E is recognized as a "consensus standard" and OSHA has said that if you follow the current version of NFPA 70E you will be in compliance.

If there is a serious arc-flash injury, there will also be lawsuits. Worker's compensation will not protect an employer if they are negligent.

OK - off my soapbox.... Sorry for the long-winded epistle.

 

[rcw retired EE]

Look at the May/June 2006 issue of IEEE Industrial Applications Magazine page 10, article "Limiting Arc Flash Exposure." The authors suggest a method of analyzing down to the MCC & 480V DP level then using a cook book approach for all feeders less than 400A. Typical device curves for each breaker/fuse size are used to calculate the minimum feeder length for each size of wire to keep the arc flash energy within the "0", "1" "2" or "3" classifications. A look up table of circuit breaker size and distance is then used for all of the smaller loads.

This sounds like a rigorous approach without having to model every load.

I am simplifying the method some, take a look at the article.

 

[dpc]

I have read it. It's basically the approach we've been using for the downstream devices. You also have to factor in the transformer size (available fault current).

If HRC goes above 2, I'd be reluctant to use this approach because this implies fault clearing is being done on the inverse part of a breaker curve or fuse curve.

With a molded case breaker instantaneous trip, it is almost impossible to exceed 8 cal/cm2 at 480 V for any realistic transformer size.

 

[JensenDrive]

Interesting, and selective modeling is a valid approach if you have a good feel for what you are doing. It has been a notable expense to enter the entire plant in SKM. About 2/3's of the breakers in the plant are instantaneous only, mainly to motors. All I am doing is confirming that the instantaneous is set low enough to work; if the instantaneous works, arc flash is always level 0, I believe. I suppose we could have left all those items out of the model. Maybe after this effort, I suppose I should talk to the customer on how to do a report where only appropriate items are analyzed, rather than the brute force method we have done.

As a side note, however, this effort forced them to do an as built of a 30+year old plant. The old one lines had tons of errors. There is another possible project in the wings to redraw/revise the plant one lines. Till then, the SKM one lines are going to be the most correct representation of much of plant.

 

[alehman]

JensenDrive - I suspect you may be correct as to most calculations stopping at the MCC level.

Given that the standards use empirical formulas that are estimates at best, and that there are frequently large uncertainties such as available fault MVA from the utility, is it necessary to construct and maintain an extremely detailed plant model?

Would it be appropriate to have a base model as rcwilson suggested and add detail as needed when work is contemplated on a particular motor or panel? However, I don't like the idea of using "typical" device curves or rules of thumb. Just add the actual branch data when needed.

In spite of the growing L/E ratio, engineering judgment must still be applied to serve the owner's best interest while adequately protecting our interests.

In regards to Zog's comment - not to pick nits, but my understanding is that OSHA will only directly reference Part 1 of NFPA 70E (2000 edition), pertaining to installation standards.

 

[Zogzog]

Alehman,

Your understanding is correct, however a few references are made to the safety related work practices in 70E and those practices will be implemented in phase II of this OSHA revision.

The way I understand it, the safe work practices from 70E implementation into an enforceable OSHA standard is a very time consuming process so they are doing it in phases. I think (hope) that when they release phase II it will be based off of the 2004 70E.

 

[dpc]

"if the instantaneous works, arc flash is always level 0"

I don't think this will be universally true. Arc-flash level will be low, but it could exceed 2 cal/cm2, depending on the available fault current. Probably safe to say it will never exceed HRC #2 (8 cal/cm2).

 

[dpc]

One other point - I assume the SKM arc-flash module is taking care of this for you, but IEEE 1584 requires calculation of arc-flash energy at both 100% and 85% of the calculated arc-flash current (below 1000V).

At 480V, the calculated arcing current (100%) is already drastically less than the bolted fault current, so 85% of that can be a surprisingly low value.

 

[JensenDrive]

Thanks for all the comments. On thing I got out of this string is an idea that while ideally all arc flash conditions should to be analyzed, but that does not mean they all need to be analyzed in tremendous detail with a computer modeling program; some form of good judgement approach can be developed where there is repetitive cases and predicable results. I am glad my company was given a free hand to model the plant (a wastewater facility) in such detail, but I can see some form of selective modeling will need to be implemented for the plants with a tighter budget.

dpc: I was basing the "always 0" on reports by SKM,using the currents and arc flash levels that SKM calculates. I was just browsing the reports, and SKM says level 0 for all the cases I looked at. I see your point that sometimes the number may be greater than 0, but that it will be uncommon.

 

[DanDel]

My approach to Arc Flash modeling has always been to provide the customer with a quote for the complete facility (if we recommend using tables for some components, then why shouldn't the customer use the tables for the entire facility?).

We always first have a consultation with the customer whereby we explain that the scope of the project can be limited to certain components, to limit the financial burden. This way, everyone knows up-front what they will be getting, and there are no liability issues as mentioned by dpc.

 

[peebee]

Re: "one could interpret any box that an elecrician could open up and work on live could have an arc flash hazard label as well". I don't buy that. I do not believe that it could be reasonably interpreted that every non-residential ballast box and receptacle and incandescent lampholder socket is supposed to have a big orange sticker on it. If the key issue is not exposed energized equipment (shock hazard), but rather interaction with equipment, then I would need to evaluate the risk every time I switched the lights on (at least on a 277 volt system). We have not yet gotten to the point where 24VDC remote controls are required for all 120 and 277 volt lights (but who knows, maybe that will be part of the next code cycle).

I do think it's reasonable to expect the hazard to be evaluated prior to working on energized equipment. I just don't necessarily think that all of these "smaller" items not mentioned by NEC 110.16 are meant to be analyzed in detail using sophisticated computer programs and permanently labeled.

Would it be safe (conservative) to say that all downstream VFD's or control panels or outlets or whatever are the same hazard as the panel feeding them? It's not clear to me if it would be possible for a downstream device to exceed the hazard of the upstream panel feeding it due to lower fault current & longer clearing time. But if so, that would certainly be an easy and reasonable way to determine the hazard.

At some (most?) sites the standard uniform provides HRC 2 protection. That sure simplifies things with most "smaller" devices.

I believe you would be in full compliance with all codes (but maybe not taking very good care of your client or electricians) if you just put a sticker on every panel, etc., stating that a hazard exists and you need to deenergize or evaluate the hazard prior to opening the device -- I don't believe the specific hazard level or flash energy density is required to be stated on the label. Acutally, I think some panel manufacturers include package such general warning stickers (loose, to comply with NEC 110.16 requirements for "field marking") with all their panels anymore.

My take on it? Provide labels EXACTLY where they are required by NEC 110.16 and you should be OK.

 

[dpc]

I think we are focusing too much just on labeling. NFPA 70E (2004) does not actually require any labels. It does require a documented hazard assessment. It also requires use of an Energized Work Permit that includes the arc hazards for the task. I've heard that the next release of NFPA-70E will require labeling, but I haven't seen it yet. But basically NFPA 70E is an electrical safety standard, not an arc-flash label standard.

The 2005 NEC requirement for arc-flash labels only applies to new equipment.

If you are hired as a consultant to perform an arc-flash study, as DanDel says you should define the scope of the study.

If you are an owner or plant engineer responsible for electrical safety (in the US), you need to accept the fact that compliance with NFPA-70E is the only thing that will keep you out of trouble if there is an arc-flash accident in your facility.

 

[alehman]

Thanks dpc. I think that post summarizes the situation well.