jmeyer wrote:
I have a couple questions that I cannot seem to get answered. We are upgrading our service switch gear in a month or so. The old gear was was in our study a little over a year ago and was found to be Extreme Danger which prompted the upgrade. I read where this upgrade requires a new study in that area. That makes perfect sense. What I am having trouble with is, if this is required for the new gear it will change ratings downstream also. Will those downstream which many are Category 0 and that category no longer exists what do I have to do with the old labeling? Do they all need removed, changed or is a full blown study needed again? I have also read that pretty much everything over 50V is now to be studied. Only things 60A and 50HP and up were done in our original study. So, with the service changes and required study of that does that force the rest of the building into a new study too. I am having a hard time getting any answer from the people that did our study and that will be doing the study for the new gear. Any help, direction, anything will be greatly appreciated.
Thanks,
Joe Meyer
Let's break this down. At a very high level we have to consider what the status of NFPA 70E is in the context of laws and regulations. NFPA 70 (National Electrical Code) references 70E. Each state adopts succeeding versions of NEC although it can be several years before this happens. NEC references NFPA 70E and each revision will reference what was at the time the current revision. So for instance the most recent edition of NEC is 2014. In my State (North Carolina) they have not yet adopted 2014 edition. That edition references NFPA 70E, 2012 edition which was the then current edition at the time. However the reference is also in the fine print notes, which are not "Code" or required by regulation but are merely guidance. Second there are no known laws or regulations of any other sort that specifically reference NFPA 70E in any specific edition. So unlike NEC, there is no "force of law" behind it.
This puts it in the status of a consensus safety standard, similar to NESC, NFPA 70B, NFPA 79, ASHRAE, and many other similar standards. So here's how it works. Say you have an arc flash incident or simply an inspection at your plant by OSHA. OSHA expects that you protect your employees against recognized hazards. How you do this is up to you. If the plant did NOTHING as in didn't even recognize arc flash then OSHA would use whatever standard they felt appropriate to apply, which would in all likelihood be 70E. But if the plant chose to implement say NESC 2012 edition as their standard for arc flash, then the OSHA inspector would have to show why NESC is invalid relative to 70E, which most won't do. They will in effect use whatever standard the plant is using so long as it is reasonable. Then it's only a matter of arguing about what the recognized engineering approach/standard is. Even if the plant was still using NFPA 70E-2004, it is unlikely that OSHA is actually going to bother with the vagaries of 2004 vs. 2015 edition.
In a similar vein, you said your equipment was rated "extremely dangerous" and worried about "Category 0" vs. some other rating. Chances are that there's a problem here. 99% of the software used to generate those labels is not compliant with ANSI Z535. Unlike 70E, OSHA does specifically reference ANSI Z535 for safety labels. The problem here is that the signal word "DANGER" is only to be used for hazards where there is an immediate and highly likely chance of serious injury or fatality. This would be applied for instance to an access door to a screw conveyor or bucket elevator, or on the access doors to live buswork on the back side of switchgear, or to an access hatch on the side of a tank full of acid. These are all cases where opening the door and going in is virtually guaranteed to be a serious injury or fatality. Everything else which has the same potential injury (serious personal injury or death) gets the WARNING signal word. ANSI Z535 makes this distinction very apparent. So in the case of energized electrical components where the door is inside the restricted approach boundary (or MAD), the shock label should be DANGER. In other cases it should be WARNING which is the case for most MCC's on the front side. In the case of arc flash though it should be WARNING only for all cases given that the likelihood of an arc flash is low although the injury is severe. So keep in mind that your equipment is 99% likely to be mislabelled.
The issue here is somewhat lost to history. There is an informational note in 70E that says that anything over 40 cal/cm^2 should be treated with increased caution. What that means is undefined. It can mean anything and so it really doesn't even belong in any safety code, even if it's only a fine print note. The original source is not clear. I've heard two myths for the existence of this fine print note. The first is that one would be killed instantly on exposure to an incident energy over 40 cal/cm^2 due to arc blast. This one is outright false. First off, arc blast is caused by air heating up inside the enclosure and being vented when the enclosure ruptures. In modelling/testing done by CIGRE, this occurs at around 1-2 cycles almost independently of enclosure size or arc energy, and the pressure is an order of magnitude less than that necessary to cause human organs to rupture. And more importantly, the pressure does not increase beyond this initial release. So overall this myth is just that...a myth.
The second myth is that at the time when the 40 cal/cm^2 suits were being produced and were made out of the same material as welding "greens" and were thus nicknamed "pickle suits", the biggest one on the market was 40 cal/cm^2. Today flame resistant PPE is available well over 100 cal/cm^2. The literal "40 cal/cm^2 barrier" no longer exists.
My personal recommendation is to pick an upper rating for incident energy. Two popular ones are 40 and 100 cal/cm^2. Personally, I'd like to set it to 12 cal/cm^2. During the training explain that PPE is only supplied by the company up to this level and beyond that, avoidance is the only valid technique. From experience working in a plant where some buses were rated at around 150 cal/cm^2 (and not being pure modelling garbage because for instance ETAP and SKM default to Lee model above 15 kV), at some point you may have to do this anyways. This means for instance turning on instantaneous overcurrent protection or using upstream equipment to disconnect power. It does not mean that death is guaranteed, which would be the case only if an arcing fault occurred near the person working on the equipment, but only that to avoid that possibility in the first place, extra STEPS to reduce the hazard are performed to reduce the hazard or the likelihood.
So moving forward....incident energy is a function of the electrode gap, distance, voltage, current, and arcing time. Except for the gear itself, changing switchgear most often only affects the arcing time because if the new breaker (or relay) trips faster then it will reduce incident energy of the downstream devices. Often the effect is insignificant beyond the next downstream device. If you know that you have reduced opening time though the labels don't necessarily have to change because at worst, the label overstates the hazard, which is harmless compared to the other case (understating it). The only two cases where switchgear can increase the hazard downstream are when it adds significant current limiting (which increases opening times) or when it increases the opening time. Since switchgear does not alter the voltage, arc gap, or working distance of downstream equipment, these are not affected.
As to a "50 V" cutoff, please be aware that there is a serious problem with modelling anything over 15 kV or under 250 V. In the case of over 15 kV, there are no tests that have been performed so it's all conjecture. However it is well known that the Lee model is invalid at these voltages. The only model that OSHA recognizes currently as valid (granted this is for generation/transmission/distribution but the same principle applies) is ArcPro. A slightly different problem occurs below 250 V. At that voltage, arcs are very erratic and are not self-sustaining for the most part. There are no nice stable arcs in most cases. Although IEEE 1584 purports to be valid down to 208 V a lot of more recent test work shows that it isn't and the entire range at those voltages is based on a single data point, not something that gives anyone strong confidence.
Due to the problems with the results, NFPA pulled a section from the Code that previously gave a range below which one could assume that everything is under 1.2 cal/cm^2 based on test work that shows that this is not always the case. IEEE 1584 still has this exception but it is likely to change (although not clear what it will change to) in the next edition (whenever that is). NESC simply has a cutoff of 4 cal/cm^2 for everything under 250 V. So any "analysis" from 50 to 250 V that is done using the most recent standards and research should simply look at the voltage and/or fault current and give a single value since no calculation method (Lee, IEEE 1584) is valid since those rely on having stable arcs. So depending on the conditions at your site, the "Category 0" labels in all likelihood would not have to be removed.
Third, you stated "under 50 HP" and "under 60A". These refer to two entirely different modelling issues. As motors are inductors during a fault, the magnetic field in the motor collapses, causing the motor to become a generator and pushing additional energy into the system. The effect depends on the size of the motors. Generally motors under 50 HP are too small to be a significant contribution so they are not modelled. Second is a little more troubling, the statement of a "60 A" load. The rated size for overcurrent protection reasons is immaterial to the modelling effort because during a fault, the current is limited only by the impedance of the system and generally not by the thermal rating of the conductors because the fault occurs over such a short period of time. The important thing to consider though again is the overcurrent protection device. The device that protects a 60 A load is likely to trip very fast and some engineers use this fact to eliminate modelling these circuits since it can be shown that except for very long conductors or large inductors it is impossible to exceed 1.2 cal/cm^2 with a 100 A, 480 V breaker. This is not true for higher voltages or for devices that take longer than about 1-2 cycles to trip.
So in summary:
1. No, no need specifically to go to 70E-2015. Whether the NFPA folks like it or not, this is a plant decision. OSHA does expect companies to stay up with the latest techniques for avoiding recognized hazards, and shock and arc flash are two of them, but as a practical matter they don't nit pick that much.
2. Whether or not the "entire" study is invalidated depends on the switchgear design and effects on downstream loads. It should be clear from inspection what this means. Relabelling at least the first layer of downstream devices is beneficial but not absolutely required. The idea that anything is labelled "EXTREME DANGER" for arc flash reasons is troubling because this is specifically a violation of ANSI Z535 which is not just a recognized consensus safety standard but is specifically referenced as a labelling standard by OSHA so there is no room for deviation here. So these labels must be changed to be OSHA (ANSI Z535) compliant under any 70E standard.
3. All equipment over 50V would need to be analyzed in general but for the most part this is probably purely inspection. I inspected a large chemical plant operation myself last year including a lab area that was all 208/120V distribution as well as numerous lighting/control systems and did the entire analysis in about a half a day once we agreed on what the basis for the analysis is.
