Something that has always troubled me regarding arc flash practices, is the apparent difference in emphasis between "Arc Containment", and "Arc Protection".
Just to set the scene, we use the mantra, "The only live working allowed is proving dead".
Successful arc containment means that if an equipment does produce an arc, personnel just outside the equipment should be safe while wearing normal cotton clothing, because the enclosure is designed, and often type tested, to ensure the arc is dealt with in a safe manner, usually by exhausting through a roof flap.
I accept that for, say, the utilities, de-energising a power supply might cause so much downstream discruption that working live is necessary. Blacking out a city is not an option. But for this, full flash PPE can be used as appropriate.
However, for the vast majority of cases, turning the equipment off before entering the enclosure is not only possible but highly advisable.
With today's technologies of portable or remote measuring devices that can log the result or feed it remotely via fibre-optic leads, I would go further to say that live measurement can virtually always be done without having to compromise the enclosure (open the door).
Why then is it not universally accepted that arc containment should be a priority, that live working be not permitted except under exceptional circumstances, and that arc protection (PPE and approach distances etc.) should come as a last resort?
De-energise, access the enclosure, prove dead, attach monitoring and measuring equipment, secure the enclosure, re-energise and work in safety, remotely. This sequence should always be the first solution.
Then the equipment, and its enclosure, can and should be designed with all the features in place to ensure this is made easier. Arc contained enclosure with arc flaps, locks on doors, preferably interlocked with the feeder breaker, indication on the outside that the equipment is energised, etc. etc..

The idea that it is acceptable to open the door of an equipment that has an arc flash hazard while it is still live should be retired to the history books!
Am I wrong?

It's kind of a solution looking for a problem. Arc resistant switchgear drastically reduces the likelihood of injury in the event of an arcing fault while the equipment is in a normal operating condition with the doors closed and latched. This is also the time when the risk of an arc flash occurring is extremely small. I have not yet found any cases in the files of companies where I have been employed of "spontaneous" arcing. There have been a few cases where breakers have exploded while being operated but there were several other safety rules that were violated leading up to this.

The most likely scenarios involve equipment which is open and being worked on. Arc resistant gear offers no protection whatsoever for this scenario. Thus I haven't wasted any dollars on installing arc resistant gear whatsoever because it offers no protection at the point when it is most needed.

I agree with Paul. The significant extra cost or arc resistant gear is hard to justify when the likelihood of an arc is when the gear is open and being worked on.

I agree with Paul! The better solution is design or modify your equipment so you remove the need to work it hot. Move all controls to a separate cabinet and use 24VDC instead of 120VAC. Install remote sensing for higher voltages to ensure the circuit dead. Remove mains from the MCC/panels. I think this would better money spent.

The comment referring to installing remote sensing...to ensure the circuit is dead... assumes that your NOT in intimate contact with the equipment when it becomes energized, or that you can remove yourself from danger in the event the circuit becomes energized. Please start including a safe grounding practice in this and all work procedures that require contact with de-energized equipment. You can also have a "close into ground" position on your switchgear provided on new installations. This option has been around for decades. In the event of a fault or short circuit, the energy is shunted to ground and may have no arcing seen or felt by the worker(s).

Hi PaulEngr, my main issue is with "working hot". I do agree that in normal operation, the risk of an arc is small, and that it is greatly enhanced if the door is open and being worked on. But that's the point. Why are poeple working on equipment when it is live, when there are always ways of doing it without puting yourself in harms way? The arc protection approach promotes a general idea that it is OK to work hot with the doors open.
I'm suggesting a containment approach that says, first, don't work live. Then, only in exceptional circumstances should you approach with doors open, and work live protected with PPE.
I was thinking about equipment connected to the breaker feed, rather than switchgear itself, but the same applies.

Hi Tish53, I would ask, why do you need to work live? If the gear is dead (and grounded - even better), there is no risk of arc.
Apart from the example I gave where blacking out a city is not an option, I would challenge any "need" to work live, by saying there is usually a way to work safely without, by making and proving dead, attaching equipment for remote monitoring, then making live with personnel safely away, either at distance or better still, protected by a panel.
I was thinking more about, say, a drive connected to switchgear, not the switcgear itself, but the same is true.

Totally agree with that Sparksmore!

I'm suggesting the following;

Don't work live, which means don't compromise the enclosure when the equipment inside is live.
If work must be conducted inside, make dead, prove dead, ground conductors, do dead working.
If the nature of the work is e.g. fault finding or commissioning, same applies, but either have designed in, or attach such provision that allows the work to be done remotely.

The problem we are now seeing is that arc fault energy is going so high that PPE really is not an option.
Why stand close to an explosion when you don't have to?

This one is pretty easy to answer. First, let's get one thing out of the way right away. Overhead utility power systems are intentionally designed to be worked on while energized. They can safely connect and disconnect jumpers, pull fuses or even change them out, all while fully energized. In fact frequently work done on circuits above 69 kV is not only done live but more and more it is being done with bare hands methods where the worker is energized at line potential because this is safer and less cumbersome to perform. This is one major reason that utilities are excluded from 70E's scope.

Two industries lead the way for electrically related fatalities: construction and utilities. Both a frequently dealing with a lot more unknowns than other industries when it comes to electrical power. That's not to say that things can't be improved here but merely points out that there is a systematic reason that these two industry groups earn more than their fair share of fatalities.

Second, you simply cannot measure currents and voltages on de-energized equipment. Now you can instrument it up all you want and do all kinds of things remotely. However at the end of the day when equipment is no longer operating normally, and thus the risk is drastically heightened, you can't really assume anything. Standard troubleshooting methodology applies which means that the approach is always to perform tests and measurements to prove what the conditions are NOT. Equipment is always live until proven dead. Voltages and currents are always unknown until proven otherwise. Equipment is always considered faulty until proven functioning correctly. In this inverted logic instead of testing for the presence of voltage, we are testing for the absence of voltage. And we are also testing the meter both before and after performing the test for absence of voltage to prove that the meter itself was not faulty. And since we can't assume anything here, that is where the need for energized work comes into play. And no, NEVER, grounding out a conductor is definitely NOT a valid means of ensuring that all conductors are de-energized.

Without even getting into grounding switch failures (I've seen those too), I'm going to illustrate the fallacies of assuming anything when it comes to troubleshooting with 4 examples of real cases that have happened. 3 of the 4 occurred at world class mining operations (#2 was a foundry). All of the personnel involved are people that I know, trust with my life, and have worked with. All of them were competent. None were fatalities so you aren't going to find them published in some fatality log on the internet. It serves no purpose to identify the companies or personnel involved so I won't.

Case #1: Electrician (that's me) around 2003 goes to troubleshoot/repair a 50 HP pump. Goes to panelboard, finds appropriately marked breaker, and locks out breaker that is marked for the pump. Indicator light on breaker (it's actually MCC) goes out. Electrician opens door to starter. Now in this particular case the electrician knew the pump wasn't running and that probably all that was wrong was some sand had jammed up the contactor (it was a sand mine) and just needed to be cleaned out but on that particular day, said electrician had to climb back down from the second story all the way through a bunch of mud and water (this was a the dewatering pump for the area) and get a forgotten meter to test the contactor. Good thing said electrician did because it so happened that the breaker was miswired and said electrician would have been severely shocked/burned by contact with a 480 V contactor which was 100% live at the time due to mismarked panelboard.

Case #2: Another electrician in another facility was working after a 14 hour day on two 2400:480 V dry transformers, probably around 330 kVA, with a rookie in tow that was scared to touch anything above 120 V. All indications externally according to those perfectly functional external voltage checks that are being recommended is that this was a case of a corroded/broken ground (single phase to ground in a system that was supposed to be grounded delta-wye). Since no 1 kV voltage detector was available, said electrician opened the panel on the 480 V side after shutting off the 2400 V disconnect and tested for absence of voltage. So far, so good. Upon opening the transformer said electrician noticed sparking when wiggling a particularly corroded tap on the transformer when checking for mechanical integrity. As it so turns out the TRANSFORMER was mismarked and the conduits were arranged in such a way up in a twisted mess in the ceiling that it was impossible to tell that they crossed one another both on the primary and secondary side of the two transformers. As a follow up it turns out that there's a simple reason for the ground fault issue. Someone had never wired up X0 to ground, but it took several years for the issue to become apparent.

Case #3: Different electrician in 2013 (last Spring) on a 4160 V junction box in a mine (4/0 cable...10,000 kVA transformer...serious available fault current here) was following current site policy which stated that one could either short a cable to ground OR test for absence of voltage. Note that site engineer argued strongly against this policy but lost the argument until this event. There were two cables running through the area so electricians got confused as to which cable was which. When electrician went to ground the shielded cable the electrician got much more than the typical arc that someone gets from grounding out shielded cable. Fortunately the ground fault system minimized the amount of damage that could have occurred to some slightly charred and melted equipment, and one soiled pair of underwear.

Case #4: This occurred prior to IEC/UL 61010 standard for meters. Contractor electricians wire up a 1 MVA 2400:480 V three phase transformer. However it's an old transformer and has wires already attached. At that time, shielded wire was not required (and almost never used) for 2400 V circuits. For whatever reason contractor electricians failed to correctly identify which leads were primary and which were secondary and wired up transformer backwards. Plant electricians subsequently began startup and checkout. Transformer appeared to be functioning correctly upon energizing. However when plant electricians tried to use a meter to measure the output voltage, the meter violently exploded and caused serious injuries to the electrician checking voltage. Note that although a CAT III or CAT IV meter would not have exploded under this same scenario, if the primary voltage had been higher, the same set of circumstances could still reoccur today.

Each of these cases shows the same basic pattern...that for whatever reason not only is the equipment not operating normally, and not only have various human errors resulted in confusion and subsequent mistakes being made, but that in the first case only rigid adherence to procedure saved my personal rear end from a potential fatal shock or arc flash, in spite of a whole lot of inconvenience. In all of these cases, short cuts such as using relying on remote indicators, and/or substituting grounding for properly testing for absence of voltage, set up all the right conditions for a potential fatality.

And in all of these cases we were dealing with a situation where the equipment was very much live. The rule to assume live until PROVEN dead is clearly the only way to handle cases 1-3. Case 4 illustrates the more general troubleshooting rule that essentially you can't ever assume anything such as assuming that the output voltage was close to 480 V. These were not poorly run facilities. They were all considered better than average or best in class in their respective industries.

You can add redundancy all you want as well. Add double, even triple redundancy. This certainly reduces the probability of a failure, potentially at great cost to either reliability or operating costs. Unfortunately even with the device from Grace Engineering or a similar device called the "VoltChek" from ABB for medium voltage, we have not made it over the hump yet to providing technology that is provably superior to the current test procedure for absence of voltage. In both cases the equipment reliability really isn't the limiting factor, it's the person performing test (or installing the test equipment). The tests themselves are extremely reliable by design. Until standards bodies and/or regulators are convinced that the system is provably more reliable all that these additional test aids will do is to add an extra supplemental layer of protection, similar to measuring absence of voltage using PT's in a safer location before testing for absence of voltage directly at the work site. It lessens the risk but does not provably lower it by a significant amount.

I agree with you and most of the comments that replied to your post. I suggest a possible approach to optimizing what measures are put in place to lower hazard is to consider all the possible future activity that may present a hazard to operators and maintenance staff. Would PPE, at any level, be required for the activity? If so, then; 1) understand the quantity of that exposure over a year, how often, how long, how many people, etc. 2) the probability of an accident, I.e. the risk during the exposure (subjective) of an accident and 3) the degree of potential severity of the exposure, how high the voltage, how much the incident energy, etc. Are the measures you are considering addressing those three factors in such a way as to reduce the overall hazard?
Obviously you can have multiple measures that compliment each other and increase the probability of a succesful outcome.
Anything that minimizes or removes the need to do live work has obvious benefit. On the assumption that accidents happen due to human mistakes it may also contribute to overall reliability by not only minimizing hazard, but by minimizing equipment damaging or product stopping events. Often the instrumentation, relays and otehr electronics that allow remote operation or simple monitorign also allow more powerful diagnostics and provide other benefits from the perspective of improved maintenance practices and reliability.
AR equipment clearly shelters operators and maintenance staff that needs not engage with the internals of the equipment in such a way that they become exposed to the hazard. I would suggest this needs more than just an AR rating, but an equipment implementation that allows full use of the potential benefits of the AR rating. Implementation of whatever minimizes the need or desire to open a panel. However, the AR rating does not protect the equipment from damage if an event does occur and may not guarantee that the panels will never need to be opened while the equipment is live. Optimizing protection so that it reacts faster and protects as well as possible minimizes hazard, and equipment damage, but does not, usually, reduce hazrd to 0. Improved protection is not incomaptible with an AR rating so implementing an AR rating probably is nto an excuse to ignore, or not try to implement the best possible protection.
In summary, I do not believe there is any one answer or measure that "ensures" "0" exposure. Different equipment, facilities, processes may be better served by different sets of design decisions. Always having power off is best but that often requires, at least, verification that is the case which often exposes one to hazard. Unluckily, doing everything right to minimize hazard and prevent exposure does not ensure that all others that preceeded you did the same, as one of the other responders very clearly described. The practice to always test before touching with appropriate trusted instrumentation and procedure is a tried and true one that seems to work. Regardless many other factors, that seems one safety measure that should be strictly adhered to.

I agree with your overall goal. De-energize, De-energize, De-energize.

We never "work" on any MCC or switchgear hot, ( such as adding a breaker or changing an MCC bucket), but you still need to test for "no power" and test the meter on a known source, and troubleshooting an MCC control circuit still puts you at some risk to the 480 and can't be accomplished with no power.

and having "arc-resistance" or "AR" switchgear or MCC's does not mitgate our risk at all and in some cases adds risk since there may be more interior panels to remove to get at the testing point for power.

As soon as you open it, to 'test for power', the gear will no longer be AR.

exactly the point i was so uneloquently trying to get across

Arc resistant switchboards should be considered as equipment protection only, and never as personnel protection.

Arc resistant switchboards could ensure emergency operations even after an internal arc. This could be beneficial for example onboard diesel electric vessels in dynamic positioning operations where loss of power could take down the whole ship, or even worse an oil rig.

Risk assessments in such cases need to have a separate assessment for the equipment and for personnel. And in some cases the equipment will and must have priority. This should be adressed in the risk assessment, and proper risk reducing measures (like arc resistant switchboards) should be identified as either personnel protection, equipment protection or even both.

If you are referring to any equipment which is rated as "arc resistant switchgear" according to the standard, then this means that a specific test was done in which cloth samples are hung in front of the switchgear and the gear is intentionally subjected to a worst case arcing fault. Then the cloth samples are analyzed to determine whether or not any significant amount of burn damage has occurred. The whole point of the test is to determine whether or not personnel operating the equipment with the doors closed and latched are adequately protected. The standard specifically allows the equipment to be destroyed in the process.

So your first statement at least partially makes no sense at all because I think you are referring to a test done specifically for personnel safety.

That being said, the standard is only valid for medium voltage switchgear so I'd agree that you should never trust arc resistant 600 V class MCC or arc resistant 600 V or lower panelboards since the standard is limited to medium voltage switchgear only. There is no such thing as an "arc resistant 480 V switchgear" assembly for instance. I don't know if the standard would apply to medium voltage ANSI rated panelboards or not since I've frankly just never used that type of gear so I don't know their rules.

That does not mean that the test could not be applied to other types of equipment and provide similar results, but there is currently no ASTM/IEEE/UL standard for this so you'd have to take the manufacturer's word on the safety of the equipment.

So if you are saying that for instance arc resistant medium voltage switchgear is equipment protection only, that is flat out incorrect. The intent of the standard is clearly personnel protection. However if you are referring to the huge amount of equipment on the market now even from major manufacturers that is being marketed as "arc resistant" but is not based on any standard, then I'd have to agree.

I am not familiar with the ASTM/IEEE/UL standards, and did not know that these are for personnel protection. That sounds weird to me.

The test I am referring to is IEC 61641: “Enclosed low-voltage switchgear and controlgear assemblies – Guide for testing under conditions of arcing due to internal fault�

[SIZE=3]The switchboard test has achieved satisfactory protection when: [/size]

[SIZE=3]1. Correctly secured doors, covers, etc. do not open [/size]

[SIZE=3]2. Parts of the switchboard which may cause a hazard do not fly off [/size]

[SIZE=3]3. Arcing does not cause holes to develop in the external part of the enclosures [/size]

[SIZE=3]4. The indicators arranged vertically do not ignite [/size]

[SIZE=3]5. The protective circuit for accessible parts of the enclosure is still effective. [/size]

[SIZE=3]6. The switchboard is capable of confining the arc to the area where it ignited [/size]

[SIZE=3]7. Emergency operation of the remaining switchboard is possible [/size]

IEEE/ANSI C37.20.7 is the standard for the U.S. Typical of many switchgear standards, IEEE writes and maintains the standard. ANSI then verbatim copies (adopts) the standard. UL then tests the equipment outright, or at least witnesses the testing if a manufacturer desires UL stamps. Generally speaking a majority of customers demand an NRTL (nationally recognized testing lab) approval similar to CE (except that CE allows self-certification by manufacturers). The whole C37 series targets medium voltage power distribution equipment.

This standard is specifically for metal enclosed switchgear. Generally in the past metal enclosed switchgear was limited to fused disconnects but frankly, you can now get anything in a metal enclosed configuration. The specification is also for medium voltage switchgear. What manufacturers have been doing is building and testing equipment to the metal enclosed arc resistant switchgear standard and then calling it arc resistant whether it is metal enclosed medium voltage switchgear or not. The only issue with this arrangement is that the consensus safety standards committee does not support for instance MCC's as "arc resistant". The testing standard is just that. So theoretically at least, the equipment meets a testing standard but may or may not actually perform as expected in service, unless in this case it is specifically medium voltage metal enclosed switchgear.

This presentation lists the specific criteria for IEEE/ANSI C37.20.7 starting on page 27:
http://www.wmea.net/Technical%20Papers/Siemens%20Arc%20Resistant%20MCC.pdf

Specifically see criteria #3 and especially #4, since both are specific to personnel safety. Note that it talks about "indicator cloth". They hang pieces of 4 oz. cotton fabric all around in front of the openings and joints in the switchgear and then test it with an arcing fault. Char is allowed but glowing/flaming is not. Although an incident energy measurement is not being used, wearing H/RC 0 (nonmeltable long sleeve shirt and pants) would protect from an injury. It is certainly a personnel protection test. Thus 70E recognizes this and lists everything at H/RC 0 with the doors closed and latched. Since the IEC standard recognizes the exact same "indicator panels", and since IEC basically works to adopt consensus standards across political boundaries, it appears that they are adopting a very similar if not identical standard.

IEC is recognized in the U.S. but not held with the same regard as in EU countries for two reasons. First, CE allows manufacturers to self-certify without any third party inspection at all, which can and has led to substandard parts from unscrupulous manufacturers. Second, there is misunderstanding about standards. In the U.S., contactors are rated according to size (kw or HP) by NEMA but the duty cycle is fixed. Under IEC you have to choose both the duty cycle and the size rating. The highest IEC duty cycle rating (20 MM cycles) exceeds the NEMA assumed duty cycle rating (10 MM cycles), but retailers of IEC components in the U.S. frequently sell the lower duty cycle components (generally 1 MM or less) as a means of lowering the price, so IEC gets a bad reputation as being inferior.

Thanks for the information. I would certainly prefer if arc resistan gear is treated as equipment protection mainly. However there can be cases were arc resistant gear can offer personnel protection as well. For instance on ships that share switchboard rooms with the engine control room, an arc resistant swtichboard could ensure that engine control room operators can work without worrying if the switchboard will be able to maintain an internal arc.

In the marine industry there is a higher focus on equipment safety than on personnel safety, for obvious reasons. However, it does not have to mean that one need to sacrifice one for the other. Proper risk assessments should lower the risk both for personnel and for the equipment. It is my belief that keeping those two risk assessments separate would lower the total risk of an arc flash. Most classification societies for shipping do not require arc flash assessments though (only Lloyds Register have requirements for arc flash assessments). After the 2010 accident onboard Queen Mary 2 most classification societies are beginning to realize that arc flash needs to be acknowledged in the entire marine industry.

The company I work for owns a ship. I've been through the code wringer with marine codes once now. It was shall we say...challenging. I hope in the future that the focus becomes more on safety and less on charging for code stamps on everything. However at least they seem to follow some kind of codes. The U.S. mining regulations are an utter joke. It is quite literally whatever whim the inspector tries to pull out of thin air and has nothing to do with any clear cut, concrete regulation. MSHA is one of the few departments of th executive branch that is self sufficient. It has become very clear that the primary goal of inspections is revenue generation, not safety. This is very sad indeed in an industry that in years past (and sometimes even in the present) lost site of priorities and placed production over the value of human life. It was only about 30-40 years ago that mining companies still reported out tons produced per fatality.