It is currently Thu Jun 04, 2020 1:56 am



Post new topic Reply to topic
Author Message
ekstra   ara
 Post subject: Do you label every disconnect switch?
PostPosted: Wed Jun 01, 2016 9:14 am 

Joined: Wed Mar 23, 2016 4:07 pm
Posts: 5
2015 NFPA 70E, Table 130.7(C)(15)(A)(a) seems to require PPE when you work on live conductors. When electricians turns off a disconnect switch, they need to do voltage testing by opening the switch up and exposing live conductors.

So by requiring PPE, a label is required on disconnect switches? If so, do you label every disconnect switch, down to 208V 30A ones?

Thank you.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Do you label every disconnect switch?
PostPosted: Wed Jun 01, 2016 10:43 am 
Sparks Level
User avatar

Joined: Tue Nov 12, 2013 6:31 am
Posts: 238
Location: Port Huron, Michigan
At my previous place of employment, we were required to put an arc flash label on every local disconnect.

Every single local disconnect in the facility had a ridiculously low incident energy, so it was pretty much a waste of resources. Once you leave the individual starter, your i.e. generally goes down to a very low number.

My current employer hasn't taken that stance, and with a lot of other priorities (including relabeling everything else with an updated "study"), I am not going to volunteer.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Do you label every disconnect switch?
PostPosted: Wed Jun 01, 2016 11:46 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
KevinK wrote:
2015 NFPA 70E, Table 130.7(C)(15)(A)(a) seems to require PPE when you work on live conductors. When electricians turns off a disconnect switch, they need to do voltage testing by opening the switch up and exposing live conductors.

So by requiring PPE, a label is required on disconnect switches? If so, do you label every disconnect switch, down to 208V 30A ones?

Thank you.


What must be labelled does not come from 70E. It comes from NEC (NFPA 70), and that is not always practiced 100% of the time. NEC requires an arc flash label but does not enforce content. 70E has content requirements but does not have regulatory authority (it's voluntary). A label that reads "Warning! Arc flash hazard may be present" meets NEC requirements though it may not meet NFPA 70E requirements. Second NEC refers to "frequent" maintenance but doesn't define it. So if you rarely do maintenance, no label required. And some areas might need frequent maintenance but don't get labels if you're not doing work while energized. For example the motor termination enclosure definitely gets maintenance on a lot of motors but most plants don't do this work energized or it happens very infrequently...so no label required.

Voltrael's statement makes an assumption that you have a combination of both a overcurrent protective device which usually either is the disconnect or has a disconnect, and a local disconnect. NEC requires a disconnect within view of a motor but there is an exception for industrial establishments which most facilities take. So if you take the exception then your plant may not have disconnects within view of the motor and if you don't take the exception then most likely the vast majority of motors will require a second local disconnect.

Keep in mind that the tables are for the most general case and may not apply to you. Let's start with the voltage check. Is it above 1,000 V? If so, then you should be using a noncontact meter. If that's the case then exactly how are you going to initiate an arcing fault from doing a noncontact test? If you read ANSI Z10 as well as the introductory material in 70E your first step should be to eliminate the hazard in the first place. PPE is the last resort. If you are below 1,000 VAC then the probes are insulated. Newer meters have very short exposures (only a couple millimeters) which are incapable of causing a phase-to-phase or phase-to-ground fault on most equipment and thus can't create an arcing fault. And in either case the noncontact meter can be handled on the end of a hot stick and the meter probes are insulated and purposely designed to prevent inadvertent elecrical contact. If the panel has no exposed conductors (everything is insulated, guarded, or isolated) as most panels are, then there is no shock hazard in the first place. So without a shock or arc flash hazard, PPE is not required. Now the table is not written for this. It is written considering the case where there is an 8" tall pile of conductive metal shavings on top of the panel, soaked in oil, and where the last person working on the panel stripped everything back 2" before terminating the conductors, and where the meter probes have the tip covers removed and is using the longer 1.5" long probes while pushing the wires aside with an uninsulated screwdriver, all while it is raining. So if you do a hazard risk assessment, you will find that your situation will be quite a bit different from the tables in 70E.

Finally and this is a follow up anticipating some details of the question. So why are you testing for absence of voltage at the disconnect? I can tell you from experience that this is not a good place to be doing this activity. The test for absence of voltage should happen at the work site at a bare minimum and only happen at the disconnect as a secondary check at best. Early on in my career I had more than one "near miss" where I tested some place other than at the equipment that I was about to perform maintenance on with disastrous results. Any time that you can't visibly see (and trace) the wiring and you make assumptions that just because it is labelled "X" that it must be connected to and control "X", you are in big trouble because once in a while the label "X" is connected to "Y" or may be connected to nothing at all. Again, this is from (bad) experience. The only value of doing anything at the disconnect is because you have a source of power to do the 70E-based test (test the meter, test the conductor, test the meter).


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Do you label every disconnect switch?
PostPosted: Wed Jun 01, 2016 3:38 pm 

Joined: Wed Mar 23, 2016 4:07 pm
Posts: 5
Hi PaulEngr,
Thanks for the info. There are two sections of 70E I'm looking at here: 130.5(D) and 130.7(C)(15)(A)(a).
Looks like 130.5(D) of 70E does mention what needs labels. But this doesn't include disconnect switch.

Quote:
Electrical equipment such as switchboards, panelboards, industrial
control panels, meter socket enclosures, and motor control centers that are in other than dwelling
units and that are likely to require examination, adjustment, servicing, or maintenance
while energized shall be field-marked with a label....


I should add that Table 130.7(C)(15)(A)(a) is part of the category method, but we are using the incident energy method. So we can argue that the table not apply here.

As far as NEC's frequency requirement, it's not very clear cut to me.

I have no influence on facility's maintenance practice. They just told me that voltage testing on the load side of disconnect switch is needed after disconnect is off, due to the possibility that switch's blade is stuck. But it seems very time consuming to put on PPE for a possible faulty switch. Also very time consuming is trying to put a label on every 30A disconnect switch.

The facility is mainly 480V and 208V. Can you test voltage without opening the disconnect?

I'm not familiar with noncontact meters. Do you have a link to it?

Thanks again.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Do you label every disconnect switch?
PostPosted: Thu Jun 02, 2016 9:58 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
KevinK wrote:
Hi PaulEngr,
Thanks for the info. There are two sections of 70E I'm looking at here: 130.5(D) and 130.7(C)(15)(A)(a).
Looks like 130.5(D) of 70E does mention what needs labels. But this doesn't include disconnect switch.

Quote:
Electrical equipment such as switchboards, panelboards, industrial
control panels, meter socket enclosures, and motor control centers that are in other than dwelling
units and that are likely to require examination, adjustment, servicing, or maintenance
while energized shall be field-marked with a label....


I should add that Table 130.7(C)(15)(A)(a) is part of the category method, but we are using the incident energy method. So we can argue that the table not apply here.

As far as NEC's frequency requirement, it's not very clear cut to me.

I have no influence on facility's maintenance practice. They just told me that voltage testing on the load side of disconnect switch is needed after disconnect is off, due to the possibility that switch's blade is stuck. But it seems very time consuming to put on PPE for a possible faulty switch. Also very time consuming is trying to put a label on every 30A disconnect switch.

The facility is mainly 480V and 208V. Can you test voltage without opening the disconnect?

I'm not familiar with noncontact meters. Do you have a link to it?

Thanks again.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Do you label every disconnect switch?
PostPosted: Mon Jun 06, 2016 4:08 am 

Joined: Mon Nov 09, 2015 7:23 am
Posts: 7
Yes. Any device that operates at 50 volts or more and meets the 70E requirements for a label (i.e. ...other than dwelling units...likely to require examination, adjustment, servicing, or maintenance while energized).


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Do you label every disconnect switch?
PostPosted: Mon Jun 06, 2016 7:14 am 
Sparks Level

Joined: Tue Jan 19, 2010 2:35 pm
Posts: 132
I asked the same question of NFPA and their answer was a non-answer. I had called to ask about labeling smaller ( 30 and 45KVA ) 480 to 120/208 transformers. I see so many of these smaller step down transformers and wanted to know if I should label or not? They didn't answer, instead they asked "why are you working on transformers that are energized?" I explained that I wasn't doing the work myself - others were and I had no control over what they may do. I thought, considering that possibility and considering that both 480 and 208 volts were right there, side by side in the same enclosure, that labeling was the right thing to do and wanted their take view on what the Standard intended? They just asked why I was working on an energized transformer? After about 15 minutes, it became clear I was not going get an answer. So I gave up.

So, to answer your question, yes, I'd label everything. My reasoning may not be justified by the various codes and standards, but it works for me. A person could claim that because a device was not labeled, be it a disconnect or transformer a fused pullout or whatever, they believed there was no hazard. We can argue the reasons why this is wrong, but if there's an injury, you're going to be asked why you labeled X but not Y? If you can't justify your decision by citing a clear and unambiguous Standard, you and probably your employer are probably in trouble.

I'd play it safe and label everything. If your manager doesn't like that approach, I'd insure he knows you're looking out for everyone's interests. The employee, the employer and yourself.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Do you label every disconnect switch?
PostPosted: Mon Jun 06, 2016 10:10 am 
Sparks Level

Joined: Sat Apr 04, 2009 9:50 pm
Posts: 104
Location: San Antonio, TX
The NFPA 70E-2015 clearly and unambiguously REQUIRES labels on 130.5 (D) under the subtitle "Equipment Labeling".

The wording of NFPA 70E-2015 requiring labels is very similar to the wording on the NEC-2014, also requiring arc flash hazard warning labels (field markings). See NEC 110.16.

I have not read in NFPA 70-2014 (NEC) and NFPA 70E-2015 the requirement that labels are only required when there is "frequent maintenance". I did a search for this term and I did not find it in the latest version of the NEC neither in the NFPA 70E-2015. Maybe it was shown in past versions.

Therefore, if you perform maintenance, even though is not frequent, NFPA 70-2014 (NEC) and NFPA 70E-2015 requires an arc flash hazard warning label.

Important Note: In these sections requiring arc flash labels in NFPA 70E-2015 and NFPA 70-2014 (NEC), there is no mention to voltage. That means that low voltage is not excluded from labels. This, I believe, should be changed. My personal opinion is that, as a minimum, we should not require arc flash labels in equipment of less than 50 V. Please do not confuse this thought with the NFPA 70E-2015 threshold of shock voltage at 50V. This is arc flash requirements, not shock.

The NEC is normally enforced in many cities and it becomes mandatory. NFPA 70E is NOT mandatory, but many corporations voluntarily adopt this standard.

With respect to the information that MUST be displayed on the labels, NFPA 70E 2015 has a detailed list of requirements. The NEC is very general and it says that the markings should "warn qualified person of potential electric arc flash hazard". It also requires that the marking (label) shall meet the requirements of Article 110.21(B) and that should be installed in a location that is clearly visible to the qualified person.

So again, to answer the question ... Should I install arc flash labels on disconnect switches?

Here is a straight answer based upon the NEC and NFPA text.

NFPA 70-2014 (NEC) Perspective:

You have to answer the following two questions:

1. Is the disconnect switch and electrical equipment such as switchboard, switchgear, panelboard, industrial control panel, meter socket enclosure, and motor control center?

2. Is the disconnect likely to require examination, adjustment, servicing, or maintenance while energized?

If you answer YES to both questions, you need an arc flash hazard warning label.

In my opinion (I have been wrong many times before) is that, even though disconnect switches are not specifically mentioned in the list, they still deserve a YES in the first question because they are equipment SUCH AS the ones mentioned.

NFPA 70E-2015 Perspective:

You have to answer the following two questions:

1. Is the disconnect switch and electrical equipment such as switchboard, panelboard, industrial control panel, meter socket enclosure, and motor control center?

2. Is the disconnect likely to require examination, adjustment, servicing, or maintenance while energized?

If you answer YES to both questions, you need an arc flash hazard warning label.

NOTE: Yes, the questions are almost identical in both, the NEC and the NFPA.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Do you label every disconnect switch?
PostPosted: Mon Jun 06, 2016 11:02 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
KevinK wrote:
Hi PaulEngr,
Thanks for the info. There are two sections of 70E I'm looking at here: 130.5(D) and 130.7(C)(15)(A)(a).
Looks like 130.5(D) of 70E does mention what needs labels. But this doesn't include disconnect switch.


70E as well as NEC (which has the same phrases) has two phrases to be aware of. The first one is that it uses the words "such as". This is an example list. The latest NEC edition added a couple more examples (and disconnects aren't in the list). The second key phrase is "likely". So if you're not likely to do any maintenance on it (see definition "working on (energized electrical equipment)" in 70E), then it won't require a label. An example is for instance the covers on a cable tray that rarely gets opened (say less than once per year). If you go to work on it while energized then you will need the risk assessment. So really what this amounts to is if you're the engineer, how many calls do you want to get for equipment that isn't labelled. If it's too many, there aren't enough labels. This is juxtaposed with if you have a standing plant rule such as "everything not labelled is under 1.2 cal/cm2". If a rule similar to this is in place, it eliminates a lot of calls but you better be very sure that everything that is over the threshold has a label. Again it could be a Code violation but it depends on the word "likely" (as in how frequent).

Quote:
The facility is mainly 480V and 208V. Can you test voltage without opening the disconnect?


Yes and no. It depends on how you test. First off both OSHA and NFPA 70E only give rules for portable testing devices. OSHA states that the test shall be "check-test" as in first check that the meter is working and then do the test. 70E goes further and requires "check-test-check". The difficulty with the OSHA requirement that the 70E requirement covers is, what happens if the meter fails during or prior to the test? However we're going after a double failure scenario here because that also assumes that the LOTO failed to remove voltage so we're looking for a vanishingly small problem that the 70E requirement covers. I'll come back to this point in a moment.

Second although neither OSHA nor NFPA 70E have anything to say about how to do the test, there are IEC standards that do. In the IEC standards they specify that the portable tool for 1,000 V or below is a voltmeter. Above 1,000 V your choices are a resistive or a capacitive (noncontact) sensor. In both portable and permanently mounted equipment they specify what the detection method has to do (I believe the standard was to reliably detect the presence of voltage down to around 10% of the nominal voltage) but for permanent sensors the field is wide open.

With portable equipment we always have a problem with 70E because we need a voltage source to test with (check-test-CHECK). With OSHA standards (and older versions of 70E) it's not such a big deal because you can always test the meter if the equipment is still powered up although again that's not always the case. But with permanently mounted equipment you simply can't meet the portable equipment 70E requirement (check-test-check). So the question is since 70E is not mandatory and working on energized equipment is certainly more hazardous, is the risk of having a failed meter greater than the risk of being in harm's way in your situation?

If you accept permanently mounted equipment, then the world opens up. Grace Engineering sells an inexpensive little device with several (redundant) LED's on it that indicates voltage status on all 3 phases plus line to ground. There are also some little blinking LED and blinking neon lamp devices. One of the most interesting (medium voltage only) devices on the market is ABB's Visivolt product that you essentially just tie wrap to the cable. There is a vastly larger variety of equipment for this use at medium voltage though where the equipment is supposed to be bolted/locked/booby trapped to prevent access in the first place.

Despite Grace's carefully worded ads, you can't really use a permanently mounted voltage detector as a substitute for a portable one under 70E if you accept the idea that "check-test-check" is required for ALL voltage tests. OSHA requires only the lower standard of "check-test". The additional 70E requirement as I mentioned earlier is chasing an extremely diminishing return. MOST of the time if this is acceptable, "check-test" can be met by a permanently mounted detector. But if there is no power present (such as after a trip) then there is no way to perform the "check" and you have to make a judgement call again as to whether for instance documenting a monthly inspection where the voltage meter is checked is sufficient for the "check-test" requirement, or fall back to the old standby portable meter in those cases. There is also the argument along the same lines that if you check the meter at the beginning and end of the shift, are these sufficient to meet the "check-test-check" requirements? There are also various versions of what has to be checked. For instance is it sufficient to check A-B phase, B-C, and A-C, or do we need to check at least one phase to ground, or all phases to ground? And if we are reading across/around a disconnect do we need to read on both sides in case of a back feeding case in which case the number of checks can double again? I've seen procedures that require as many as 12 voltage checks for a single "test for absence of voltage" when this is taken to the extreme, but many electricians I've observed skip even the phase-to-ground test and only do 3 checks phase-to-phase.

As to reliability...permanently mounted equipment can and does fail. Most of the manufacturers recommend that there is some kind of redundancy or self-checking built into the equipment, such as ABB recommends mount 2 of their devices per cable and Grace has 2 lights per phase. Connections to the equipment are generally not an issue and it can be routinely inspected so most of the reliability issues that occur with portable equipment are less of an issue with the permanently mounted versions.

Quote:
I'm not familiar with noncontact meters. Do you have a link to it?


The most popular medium voltage one that I've seen is the one Salisbury sells. Fluke among others sells a low voltage one. HOWEVER all of the reputable manufacturers for the Fluke ones specifically state that it can't be used for safety purposes which defeats what you are trying to do. The second issue in general has to do with how capacitive noncontact voltage sensors work. They rely on the electrical field around the wiring to detect them. This has several problems:
1. Shielding totally disables the device. So any time you have a "Faraday cage" situation such as a metal enclosed panel, it won't work at all.
2. Electric fields cancel each other. So if you have triplexed 3-phase wiring it can be very difficult to get a signal.
3. The device only works on AC. So a big capacitor is completely undetectable.
4. They are subject too false positives as well since it just senses fields. So if there are several cables nearby you can't really tell which one is responsible. And even if you can isolate it if there are several cables running close together it can sense the inductive pickup as well.
5. The vast majority of these devices do not sense down to 50 V. And even if they do since it is all a relative reading anyways especially at low voltage, there is no way to tell whether or not there is voltage present or if it's just yet another false negative.
6. And for all these reasons (false negatives and false positives) unless you can positively prove that the cable is dead, why woul you even consider risking it?

At medium voltage the signal is much clearer and stronger and standard practice whenever possible is usually to hold the detector up to a line that is still energized and then to the line that is going to be disconnected, disconnect power, and then retest again on both lines to positively prove that the meter was working and is still working. Of course there's not always a source nearby...

And the voltmeter isn't a "proven" thing either. Even if I get a positive "zero" reading, especially if I question the ground, I will normally hold the probe to the ground, flip the meter over to reading ohms (after testing to make sure the circuit is dead!), and then test using another ground in the panel to prove that I had good contact with the ground.

All that being said, I would strongly discourage using a disconnect as a point to test from! Eventually at some point you will get a surprise. Here are three real world cases:
1. Electrician is working on bin vibrator in a plant that runs off 220 power. Electrician pulls the plug from the nearby receptacle. The enclosure is plastic so there is no practical way to test for absence of voltage. There are two cables running out of the box. While opening the box to work on it, electrician receives a very nasty jolt and finds that the cord and plug were disconnected and the whole thing was hardwired. The electrician who did the original work said he didn't have a hole plug so left the cord on.
2. Electrician is working on a screen plant with a well for spray water. Electrician locates the MCC bucket with a phenolic label "well" and then slogs through 15 feet of mud and slop and up two ladders to reach the remote starter for the pump (runs on a float valve). Electrician opens the panel and notices that the contactor is still pulled in or possibly jammed. Electrician realizes he forgot his meter. Electrician then makes the trip all the way down to the truck and back again only to find out that the starter is still energized. Investigation later reveals that the bucket had failed. Another electrician stripped the bucket but left the label on and wrote in black marker "well" on the new bucket and wrote "not used" on the original bucket. All the marker had faded away over time and the phenolic labels were not updated.
3. Water treatment plant at an industrial site appears to have a system grounding problem and a ground fault (phase-to-ground reads 480-480-0). Electrician locates the 2300:480 transformer (delta-wye), 2300 V fused disconnect, and 480 V fused disconnect, all located in a room near main power distribution that was repurposed over time and has a massive amount of both live and dead conduits at the ceiling level criss-crossing the room. Both disconnects and the transformer are clearly labelled. There are two 2300 V disconnects, two 480 V disconnects, and two transformers (identical) located in the same room. although the arrangement is "L" shaped (transformers face 480 V disconnects but are sideways to the 2300 V disconnects) the labelling is clear and the conduit appears to go in the right direction. One transformer was labelled "water treatment" and the other was unlabelled. Electrician opens both disconnects then uses a Salisbury voltage detector on the 2300 V side and a multimeter on the 480 V side as well as receiving a call from the water treatment plant operator that the lights are out. Electrician removes door on dry transformer and inspects to find that as expected, the transformer ground strap to X0 was corroded and broken. Electrician also notices that one of the transformer taps is badly corroded and reaches out to wiggle it to verify. Upon doing so the tap arcs and jumps. After further investigation it was found that in the past one of the two transformers had failed. Electricians moved wiring from one transformer to the other to keep the plant operating until the replacement arrived. When the replacement arrived electricians wired up the replacement transformer but did not swap anything back (wiring remained swapped). So the unlabeled transformer was actually the correct one.

I should mention that although these cases span about 15 years of time and they all happened to a single electrician and that I was the would-be victim in all three cases. I still tell these stories because they emphasize not only the importance of testing for absence of voltage but how to do it properly.

These are all in plants that generally speaking are considered the best in their respective industries and not some make-shift fly-by-night operations, in capital-intensive businesses where spending on safety is generally not a issue. All 3 cases show blatant problems with "proper maintenance" but also highlight was typically can and does happen even in better maintained equipment. And most importantly it highlights that although in case #2 opening and testing the disconnect probably would have discovered the missing equipment, testing at the disconnect was actually done in case #3 and would not have prevented the issue.

Instead of testing at the disconnect, or in addition to testing at the disconnect, always, always, always test at the work site, before you stick your fingers anywhere. Even if it is a blatantly obvious case such as a plastic control box with a cord coming out of it, check anyways. The whole point of testing for absence of voltage is that no matter how well maintained your plant is, no matter how well it is labelled, no matter how organized everything is, no matter how "obvious" it is, there are always things that go wrong when working with electrical equipment. If you want to test at the starter bucket/disconnect just because you've run into cases where the breaker welded shut or a disconnect blade welded shut, by all means go right ahead and do that if you can do so safely. This is an OPTIONAL check just as NFPA 70E's requirement to visually verify that the blades are open if you can see them is optional (it doesn't mandate visible blade disconnects, just that you check if they are visible). But this is totally optional. Testing at the work site, at the location where your hands are about to go, is NOT optional. And this is not just a label issue as case #1 points out. I've seen wiring that was damaged and out of place, various versions of "backfeeding", and many other situations. LOTO fails infrequently but it does fail for all kinds of reasons, but the voltage test works every time I've used it.

In case 1 and 3 above, it was only pure luck that saved me from becoming a safety statistic. In case 2, it probably saved my life. I was convinced that the contactor had either gotten some sand in it and mechanically jammed or it had welded shut and either way I was going to have to take it apart to free or replace it. I already had the tools in my pouch to do this. I actually thought twice about not going to get the meter. I was trying to decide what the chance was that the contactor was actually energized since I was absolutely sure I had everything locked out, and I did not want to trudge through knee deep mud twice more for nothing. I'm glad I did.

Coming full circle on this:
1. A permanently mounted voltage detector can positively identify whether or not a disconnect is powered off. But that's only the first step and as mentioned previously its optional anyways, unless the disconnect is where the work is to be performed. In this case it is a question of whether or not permanently mounted detectors are considered acceptable compared to portable detectors.
2. A check at the work site is always required whether or not the disconnect is checked in the first place. Arc flash by the way will be much lower here compared to at the power distribution point especially if the disconnects are fused.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Do you label every disconnect switch?
PostPosted: Mon Jun 06, 2016 12:16 pm 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
RECS wrote:
Therefore, if you perform maintenance, even though is not frequent, NFPA 70-2014 (NEC) and NFPA 70E-2015 requires an arc flash hazard warning label.


The word is likely. Likely implies frequency and/or probability. Since we do not exist in a moment in time it is always frequency. Even if the risk is 1 in a million given enough time and enough people exposed to the hazard eventually at least one will be exposed to it.

Quote:
Important Note: In these sections requiring arc flash labels in NFPA 70E-2015 and NFPA 70-2014 (NEC), there is no mention to voltage. That means that low voltage is not excluded from labels. This, I believe, should be changed. My personal opinion is that, as a minimum, we should not require arc flash labels in equipment of less than 50 V. Please do not confuse this thought with the NFPA 70E-2015 threshold of shock voltage at 50V. This is arc flash requirements, not shock.


No "bottom end" has been established for arc flash but many folks such as yourself have pretty much settled on 50 V in absence of anything else. What doesn't help is that when you read 70E for guidance even at 50 Volts there is hand waving going on. The 50 V section under 130.1 ends with "...it is determined that there will be no increased exposure to electrical burns or to explosion due to electric arcs". Expose me but exactly how in the world are we supposed to do that? This is quite literally directing the end user of the standard to "prove a negative". As an example, the IARC maintains 4 lists of things that definitely do not cause cancer, might cause cancer, probably cause cancer, and definitely cause cancer. There is only one substance on the "not" list but hundreds of things on the other lists and many of the ones on the "might" and "probably" lists would surprise you but IARC's standard for "definitely not" is in an absolute sense, something that is provably almost impossible to do. So if we use absolute terms then basically there is no "50 Volt" rule because the rule that we have is impossible to actually use in practice.

It can be proven with theoretical research that an arc cannot start (but can be sustained) below around 28 VDC by taking the DC equations for arcing and run them to the limit (as available fault current becomes infinite). If we assume infinite available fault current and look at the voltage as we reduce the gap towards zero using Hertha Ayrton's equation for instance, everything drops out at around 28 VDC. Even at very high available fault currents the voltage increases to around 50-100 V. Keep in mind here that once it starts there can be enough heat to keep restriking and with a DC arc (such as with a welder) the arc is constant so once it is established it won't extinguish, so gap plays a factor especially if we try to consider moving conductor cases such as if a conductor breaks free and strikes the enclosure or another phase.

Experiments at Kinetrics have shown that arcs can't be sustained at 20 kA with 130 VDC and a 1/4" gap for more than 0.8 seconds which works out to right around 1.2 cal/cm2 at 18". Even though this is a DC result it clearly makes the case for 120 VAC. There was some testing done about a decade ago at PS&E that seemed to suggest that below 250 V arcing faults were at worst under 1.2 cal/cm2 but more recent testing has disproven this. The biggest issue is that PS&E used the IEEE standard method at the time with a 14 gauge "fuse" wire. This is too big and for low voltages the "fuse" wire needs to be smaller. Similarly more recent testing has shown that the "box-barrier" test can maintain an arc at much lower conditions. As of right now IEEE 1584 gives some vague guidance that at 208 V or less with a 125 kVA transformer or smaller, there is no appreciable arc flash hazard. I believe that in the next couple years that when IEEE 1584 itself gets revised, we can expect the cutoff to be revised downwards and contain a current limitation (hopefully with a voltage limitation). IEEE C2 (NESC) extends this up to any equipment rated 250 V or less with 4 cal/cm2 PPE. The actual IEEE 1584 test database includes a single case at 208 V. All other tests failed (could not maintain an arc) at that time. That's all that's out there that I'm aware of for the low voltage case.

So my own feeling is that based on Ayrton's equation we can rule out anything below 28 VDC and probably anything below 50 VDC especially when we consider the years of service with high available fault currents in the telecom industry with very large UPS's all running with a nominal voltage of 48 VDC with few incidents. In fact we can capture a whole range of voltages and currents with Ayrton's equation although most of them are so low that they don't apply to very many cases. Based on the Kinetrics test we can rule out 120 VAC/DC for any current up to 20 kA. Above this voltage the best we can do is using the IEEE 1584 suggestion (<125 kVA, <=208 VAC) although we can extrapolate somewhat the short circuit current based on available transformers at that size. Above those limits and below 250 VAC (with infinite available fault current), we can reference IEEE C2 for a 4 cal/cm2 limit. Realistically looking at the sources for the IEEE C2 information (mostly EPRI), there is still an outside limit of around 20-30 kA for the higher voltages where they were unable to achieve about around 3.2-3.8 cal/cm2.

Getting back to the frequency and likelihood, working in absolutes makes sense and seems like a "safe" approach but that's not what the standard is based on. In several places NFPA 70E seems to have a precise an absolute standard such as stating where a hazard exists or where it does not exist. The arc flash tables use "yes" and "no" for risk. This seems like a satisfactory answer. However the world is not black-and-white and neither is the standard. Perhaps a better choice of words would be "likely" and "unlikely" but that's not how the existing standards are constructed. This is not just a 70E issue...it's throughout all of the NFPA Codes. From the Handbook under arc flash hazard: "An arc flash hazard exists if a person is or might be exposed to a significant thermal hazard". Similarly the definition of shock hazard refers to "inadvertent" or "accidental contact. Stepping away from electrical hazards, hazardous locations exist only if the conditions are likely to exist. Otherwise, the location is not hazardous and we can use standard wiring methods. Throughout these definitions there is a thread that probability or frequency is intimately intertwined with the "existence" or "absence" of a hazard. NFPA recognizes with all these cases that there is no such thing as an absolute iron clad proof beyond a shadow of a doubt that a hazard exists or does not exist. Although I would think that in certain cases we can prove that an electrical hazard does not exist in the absolute sense, that's not the direction that the standard takes. NFPA standards use absolute terms (yes/no, exists/does not exist) but they are predicated on a frequency or likelihood of occurrence. The labeling standard is no different since once again, the word likely appears in the definition. Once we move away from the definition, the rest of the document is a black-and-white "yes/no" condition on whether or not the hazard "exists". The reader is assumed to know to interpret this as "likely" or "unlikely".

Finally there is good reason on NFPA's part not to even go so far as to define a precise threshold. In the ideal world the risk threshold should be such that electrical hazards are no more likely to occur than any other work place hazard relative to their severity. This would necessarily be company and/or industry specific. The standard would be much greater in the nuclear power generation or airline industry compared to say forestry or commercial fishing because the hazards are different. For this very reason all of the risk assessment standards that exist are relative standards. The risk that is acceptable varies from one industry to another.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Do you label every disconnect switch?
PostPosted: Mon Jun 06, 2016 1:53 pm 

Joined: Wed Mar 23, 2016 4:07 pm
Posts: 5
Thank you for very thoughtful replies.

When you mentioned 250V 20kA, it is the arc fault current, not the short current, right? If you are doing calculations using software and you have the arc fault current, then you've done all the work to calculate it and may as well print the label because the software knows the incident energy already.

I'll keep in mind the work site testing. Thanks.


Top
 Profile Send private message  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 11 posts ] 

All times are UTC - 7 hours


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
© 2019 Arcflash Forum / Brainfiller, Inc. | P.O. Box 12024 | Scottsdale, AZ 85267 USA | 800-874-8883