Without delving into NFPA 70E or CSA Z462-08 - who thinks normal operation of equipment includes opening and closing disconnecting means? Let's assume the equipment is in good mechanical condition with no exposed electrical parts. For those who think Arc Flash PPE is required: why is it required? No references to legislation please, just good scientific reasoning.

This has been discussed many times, try searching and you will find a lot of ideas. Regardless of what the 70E says, and how different people interpret it, here are the facts.

Fact#1. Switchgear, diconnects, MCC's, enclosures, etc. Unless they are specifically listed as "Arc rated" are not and were never designed or tested to contain an arc flash and protect the operators from being injured.

Fact #2. People have and will be killed and injured operting such equipment.

Operating equipment in the normal way, without PPE is a risk some people consider low enough that they are willing to take that risk, some of those people in the past have regretted that decision. Some of the parents of those people buried thier child. I don't care if the odds are 1 in a million, I am weaing PPE or better yet switching remotely.

Opening Circuits under full load would be considered hazardous to me.Qualification of the operator would have to be considered as well as design of the system ,all considered I feel that wearing the proper PPE would be the only safe option.

Operating breakers, switches and MCC's from outside the arc flash boundary is the safest option. Standing in the boundary waering PPE would be the next best option.

In order to consider this within the designed intentions, you have to KNOW that there are no loose connections or loose material in the switch or circuit breaker. You also have to KNOW that there are no loose connections or loose material in the load being switched.

I do know of one case, and I talked to the electrician at the plant, where a breaker was closed, remotely (thank goodness) to start a motor that had just been serviced. Because of a small fault on the motor that was undetected by the service crew, when the breaker was closed in, the arc flash blew the door off the 6900V breaker cabinet, and the electrician, who was standing at the end of the switchgear lineup, witnessed a ball of fire roll out for ~20 feet.

This may not be scientific enough, but it convinced me.

Ball of Fire

I am interested to know if an analysis was done on that piece of gear, and if so, what the Flash Protection boundary was?

The analysis showed ~95 cal/cm at 36 inches, if I remember correctly.
The boundary was much, much further than the 20 feet witnessed. The door could account for part of this, since it was closed to start with, but this is another reason why I don't trust the calculated boundaries at large energy levels.

WDeanN: Was it really the worst case which could occur?

If I recall Art 130 correctly, its states that FR clothing is only required for operation on dead front metal enclosed devices that are switch substantial energy. Which is latter defines as Switchgear. So while FR would always be the safer way to go, an exact application of Art 130 would not require it for typical safety switches, assuming they are closed.

It probably was not, but I feel that there are other errors in the methods used by the 1584 in the initial equations.
On the calculation of the arc flash boundary, the majority of the tests were low energy tests with short clearing times, and the probes were all placed within ~4 feet of the equipment.
They then took this data, and after noting that there is a difference for and open event and a closed (box) event, extrapolated the data from the closed event out to all events in a box, including large energy events and 300 foot!? boundaries, where the focusing effect of the box would have little to no effect on the arc flash and blast.

Hazards while operating equipment.

Durring training I normally tell people that they can wear one level less PPE then labeled, as long as there are no exposed energized parts and the barriers to those parts are secure. If it's category 0, all that's required is safety glasses and ear plugs. If it's category 1, there may be enough heat that still comes out through cracks between breakers, etc that the long sleeve shirt and leather gloves ("when applicable", which would be here seeing no work is being done on or around exposed energized parts) should be sufficient protection. If it's category 2 or higher, then the blast potential is enough to still need the flame resistant clothing and the face/head protection, even though it is a lower hazard level (thus PPE level minus one only). If it's 40+ cal, they need to wear full rated PPE while operating (to be done only when absolutely neccessary) and should never expose that equipment. Though I agree with remotely operating, but in most cases (especially lower equipment) it's not possible. Plus, because of how equipment should be energized top to bottom, re-energizing equipment normally must be done energized (not under load though, of course). Also, every time I do switchgear testing, the power company wants all mains open first, can't do that without tripping them open energized. Also, body possition and distance is still very important (something that people have recognized, even before sep '97)

Breaker Adjudtments

Where would everyone classify adjusting breaker settings?

How can you justify that?

What is the purpose of reading the labels provided if you would systematically train persons to consider them wrong?

If it’s any consolation Bruce JAX, you are not alone. Our corporate standard allows operators to operate 480V MCC breakers in HRC 0 provided the load is down and the door is latched with every latch installed by the manufacturer of the equipment. An addition of an 8cal/cm2 smock is added to this requirement at our local site. I realize many on this forum will disagree with that approach. The only thing inconsistent with the disagreement is that most of these same people will act like it is acceptable to reduce the PPE based on task when using the tables; however, with a calculated number this is no longer acceptable. Do I understand the difference from a litigation standpoint? - Yes Do I understand the difference if the concern is worker safety? - No The people who wrote the tables are smart enough to know that clearing time and arcing current do not change if the door is closed. As for me, I do believe that there must be some reduction in hazard when the door or cover is securely in place. Ref. 2012 70E 130.7(C) (15) second paragraph of Informational Note No. 1 This note is either applicable to the likely hood of or the results of an arc flash event or it is wrong. How can you have it both ways?

Originally Posted by BruceJAX View Post
Durring training I normally tell people that they can wear one level less PPE then labeled, as long as there are no exposed energized parts and the barriers to those parts are secure. . .

BruceJAX I would question what you are doing by personally making the decision to train people to your belief that you can drop down a category of PPE if there are no exposed parts and they are secure. Have you done a detailed hazard analysis that shows the arc flash energy will be reduced to the proper calorie requirement to drop to a lesser category of PPE? As a customer, if I hired you to conduct this training and we followed your protocol and someone was involved in an Arc Flash accident that caused burns greater than second degree then my company's lawyers would be contacting you. 70E PLAINLY states the requirements for the situation you posed with no gray area. Even though you might not agree with them, I have seen and guarantee that a lawyer will ask why the company did not follow the information on the label. I hope you have a tremendous non-liability claus in your contract.

I think that we can all agree that the doors on the equipment provide SOME level of protection even in non-arc resistant gear. But there are two factors to consider. First, it strongly depends on the amount of arcing fault energy released. Beyond a certain pressure level (the precise level of course being uncertain), the doors will get ripped off and it doesn't matter anyways. AIC claims notwithstanding because those are testing with bolted faults and not arcing faults. The problem with both that threshold and the energy reduction is that since there's no data available, you can't quantify it.

The idea of a "1 level reduction" seems simple but remember, the "levels" are arbitrary and not on a geometric or linear scale. So the "jump" from one level to the next kind of screws up the whole "1 level reduction" idea, regardless of whether there is any kind of scientific rationale for it or not.

Second issue: risk. Major chemical plants make this estimate all the time. Some things are simply inherently risky. We've all read the accident statistics about auto vs. airline flights. Somewhere, someone is always making a risk assessment even if they don't intend on doing it. We all roll the dice every day, and some of us lose the bet. It's a simple fact of life that it is physically impossible to protect against EVERY situation. The question that you have to ask yourself though is to weigh one situation against another and decide if the risk is acceptable or not.

In the case of switching or pretty much any activity involving any equipment (electrical or not), we are supposed to weigh the odds. OSHA demands this pretty clearly. Usually, the decision is pretty easy. Sometimes though costs, comfort/convenience, time, and other factors make it such that we really need to consider whether a particular task is "safe enough" or not, since as I just said, nothing is totally safe.

So when it comes to electrical tasks like switching, the question is whether the task is "safe enough" that PPE is required, or not. If you never, ever make the evaluation (which you are required to do these days), that is in itself a decision...and carries some risk.

We don't really have to question this though too much. There are at least two data sources that I'm aware of. The IEEE Gold Book contains a wealth of failure rate data on almost every piece of electrical equipment. In this particular case one could use the failure rates of circuit breakers and in particular estimate the failure to open on demand or to open too slowly, since the other failure rates are pretty much inconsequential. I've done this approach and it pretty much agrees with the second approach. The second one is that the NFPA 70E Technical Committee with the caveat of doing proper maintenance (by the way, the Gold Book data also produces the same result), gives specific tasks that they feel have a low enough statistical chance that it is an acceptable risk.

The one problem that I see with the IEEE data set is that it is extremely dated. Most of the data is from the early 1970's and 1980's. For instance it shows that vaccuum contactors are MORE prone to failures than air contactors, which is certainly not true. It also shows high failure rates for trip mechanisms. This is certainly true of the old electro-mechanical ones but the newer microprocessor based trip units have in my experience held up much better. Even the microprocessor based overload relays are a lot better than the old eutectic alloy ones. They're all getting close to the point of having the reliability of fuses which have zip, zero, zilch probability of failure to trip on demand (only nuisance trips can happen from what I've observed).

The other arguments such as doing switching while unloaded are certainly valid. It certainly raises the potential failure rate but perusal of the IEEE data doesn't quite give enough information to quantify it to this level.

There is also a wealth of failure data for electronic components out there especially some of the military publications. But, when it comes to electrical gear, I was not able to ever find enough data to quantify the equipment down to the electro-mechanical failure modes that could occur.

With this in mind, what you certainly cannot do is to MIX the statistical probability of an incident with the consequence. We have a binary outcome on the consequence side. So multiplying the "PPE level" or incident energy level by the probability doesn't work. You simply determine the tolerable risk level (acceptable probability), whatever it might be. If the actual chance of an arc flash resulting in an injury is less than that threshold, PPE is not necessary. Otherwise, either equipment/task redesign is necessary, or as an ultimate fall back strategy, PPE is required.

It is certainly possible to do a risk/hazard analysis to determine different HRC levels. There is an annex in NFPA 70E to assist in this effort. It cannot be an arbitrary reduction of one level, however. If you have a consulting engineer do an arc hazard analysis, it isn't inexpensive. Adding a risk/hazard analysis will make it even more expensive; this is not normally included in the scope of work of an arc hazard analysis.

Be careful there. You can't do a risk/hazard analysis at different PPE levels. You either have sufficient PPE to avoid one particular hazard (potential for a second degree burn) or NOT. Current arc flash analysis statistics tell you nothing else.

It is possible to estimate equipment reliability...the chance of having an arcing failure with any particular task. If that's below your risk threshold (or the 70E Technical Committee's), then PPE is not required. Otherwise it is and the incident energy calculations tell you how much.

If you are doing something to reduce the energy (maintenance switches, increasing distances, etc.), then PPE reduction is certainly possible for those instances where reduced PPE is possible. Otherwise, it's not.

The key to understanding risk assessments is that you first do the analysis and determine whether the risk is acceptable with NO PPE or any other protective measures. Then you attempt to engineer out the risk in some way. If you cannot lower either the probability of occurrence or the consequence, then you resort to PPE to mitigate the risk. If the incident energy level is above 40 cal/cm^2, then PPE is not an option and other methods to address the hazard MUST be employed.

But didn't the technical committee do just that to develop the task tables? They lowered the category by one or two depending on the risk that a certain task would create an arc hazard.

You might find historic data on the probability of equipment failure, but this isn't the same as the probability of an arc being caused by opening a door or human error when testing voltage.

Yes. Care to figure out how to duplicate their results? Or explain the justification? There doesn't seem to be any forthcoming.



You have to consider what percentage of those failures were because of opening a door for instance. It gets a little arbitrary but if you don't feel secure with this, just stay with the data as-is. Another problem with reliability data is that you would be interested in dangerous failures, not all failures. The reliability data would include failure to close when required for instance, which is obviously a large fraction of the total failures. On top of that IEEE Gold book data is very, very dated and hails from a time where for instance vacuum contactors were less reliable than arc contactors in medium voltage.

Human performance difficulties are well studied and there is lots of data on this. I like to use HEART because it's fast and easy. For more of a procedure/task oriented approach there are better data sources. These studies come from various military organizations who are both keenly interested in human performance difficulties, and have ample funding to do a good job on studying this. The data for instance shows that the classic E-Stop button in those circumstances it's intended for have at BEST a 30% chance of being used correctly.