Hello,

Brand new to this forum and to Arc Flash. I am a relatively new engineer working on some Arc Flash evaluation. First, let me say that my work will be evaluated by a professional engineer before it is accepted as correct. So don't worry if I sound confused about this. I just don't have access to this professional engineer for a while.

I have a couple of questions.

1.) I keep reading that you should exclude the main breaker on a panel if it is not in a separate compartment because plasma generated during an event could cause an electrical path from the load side to the line side, rendering the main circuit breaker useless. So you are only supposed to consider protective devices upstream of that panel. We are trying to minimize the amount of information we have to get from the job site. Does the above recommendation mean that we can forget about collecting main circuit breaker data from the furthest panel down the line? I figure that since it doesn't feed anything that is being arc flashed,there is technically nothing downstream of it so we don't need any feeder breakers from it. If we are excluding main circuit breakers then theoretically we need nothing except the conductor size and length to that panel, but no breaker info (excluding upstream protective devices). Is this correct?

2.) We are excluding smaller 208/120 panels in the system and labeling them with a generic label. In my mind this means that the entire load on the system will not be accurate and the software we are using (SKM PTW) will not get the correct results for the remaining panels that are being calculated. My reasoning is that the total available fault current from the utility is reduced as your system grows since running loads are essentially controlled short circuits. With many pieces of equipment in parallel, there is less available fault current from the utility because some is being used. So it seems that even though they are being labeled with a generic label, they must still be included in the model so that the current they draw is accounted for. But, doing this seems to negate the time savings of using generic labels since they have to be modeled anyway.

Also, IEEE exemptions covering panels less than 240V and XFMRS < 125KVA seems to cause this same problem. Wouldn't leaving items off of the model affect the available fault current?


The only thing I can think of is that during the incident energy calculation, all parallel loads are assumed to be not drawing any current and thus the full capacity of the utility is available to a single series connection of panels and xfmr's all the way down to the last panel downstream. Is this the assumption? Otherwise, how could you accurately account for the normal day to day power consumption of your system without accounting for every light and fridge attached to every panel?

Hi and welcome to the forum!!
1). You are correct. The data collected from the OC device in the most downstream panel will not affect the arc flash calculations. We go ahead and collect this data just for completeness of the one-lines.

2). Your assumption is correct. The full capacity of the utility is assumed to be traveling through the conductors to the location of the fault.

Thank you for your prompt reply. This is very interesting. So for a standard panel that is not compartmentalized, the arc flash calculation for that panel is calculated without considering the protective device on that panel. This means that even though technically you will have the option to have a load and line side calculation in SKM PTW, the safe bet is to apply a label with only the line side calculation. Is it ever common practice to put two labels on standard 20x6 panels? This seems excessive and that just using the line side label would be practical. Am I on the right track?

I have never put two labels on a single non-compartmentalized panel.
My customers usually want only the worst case label. Modifications to the PPE, on the label, are done through their EEWP.

In an MCC with a main circuit breaker, you still typically see two labels for the whole panel. The first is the rating for the main circuit breaker compartment (which is protected by the upstream device). The second is esssentially the arc flash hazard from the bus, protected by the main circuit breaker.

Typically once you get downstream of the breaker though at the utilization end of things, the incident energy is low enough that you can ASSUME 1.2 cal/cm^2 or less. It is typical to use that assumption so that you don't have to go crazy labeling everything down to the motor level. However once in a while you will find downstream equipment that is big enough to matter. At this point you have two choices. You can either accept the worst case (meaning that the breaker load side is treated the same as the high side) or you can go ahead and label on the outgoing side as well.

Do NOT assume by the way that the further you get from the transformer, the lower the incident energy. This is definitely not always true. The actual incident energy follows something like a parabola. As you get further away the arcing current usually falls off but often this results in increased trip times, and increased overall incident energy. With resistance grounded systems in particular cable capacitance eventually (with very long lines) increases the current at the fault (but not at the resistor) as it begins to look more and more like an ungrounded system. If you have a few fairly long cables, it is frequently very helpful to model them anyway just to check for this possibility.

I think what Paulengr is referring to is ferroresonance which is caused by long cable run capacitance. This isn't a problem in building distribution systems and doesn't contribute fault currents.

Thank you to everyone for your input. I definitely know where to go if I have any further Arc Flash questions.

It's not ferroresonance. True ferroresonance has to do with nonlinear effects of transformers when you saturate the core (when your perfect inductor is no longer perfect). Most of the time though the resonant point of an LC circuit gets attributed to ferroresonance, but typically this only results in subtransients that are minor players in an arc flash calculation.

The problem is the interaction between overcurrent protective devices and decreasing fault currents. Most overcurrent protective devices have an inverse time relationship so the trip time increases along an exponential as the fault current decreases. So even though the arc power is decreasing as the fault current decreases due to the increase in impedance (which could be pure resistance for this exercise), and even though the IEEE 1584 equation has an exponential associated with it too, the rate of change is different than that of the increase in trip time. When you ratio the two, the effect is that the result (incident energy), can either decrease or INCREASE and typically tends to increase at least initially. Try it some time in your choice of software...dial up the cable distance over a few points and see what happens. Sometimes it does the expected thing but quite often it does just the opposite.

Most people assume that relatively long buses keep incident energy low and generally they do, but do not assume that this is always the case because of the effect on extending trip times.

Paul,
Why would the MCC be different than a switchboard or a switchgear and have two labels? I would think that the "plasma effect" would apply as well. Four years back we used to differentiate between the line and a load side of the main MCC breaker, but after extended going back and forth with Square-D and GE we changed our thinking. The manufacturers do not test their MCC section separation walls to withstand an AF and they recommend using line side AF rating for labeling for safety.

Thanks,
Marek

In a panelboard, the main breaker is not isolated from any other components. It would be invalid to rate anything in that panel lower than the line side of the main breaker.

In an MCC on the other hand, if I'm working on a bucket downstream of the main breaker, I would have to trigger an arcing fault on the line side of the main breaker in order to cause an arc flash of a magnitude equal to the hazard posed by the line side of the bus in the main breaker panel. Otherwise the main breaker has a chance to arrest the arcing fault and must be considered for the opening time of the overcurrent protection. This gives me a much lower arc flash hazard potential when working on energized MCC cubicles compared to working on the main breaker (if equipped) inside the MCC.

The mistake that is quite often made is to ignore the line side bus within a given MCC cubicle. Placing two labels on buckets other than the main breaker would serve the purpose of indicating what the maximum arc flash hazard is from the cable/motor that the MCC is operating. Within the MCC bucket itself, the line side (or load of the main CB) is the correct rating.

Paul,

I understand what you are saying, but again the MCC or a Switchboard is an open concept gear and the isolation is not considered to be an substantial barrier. In most cased for switchboards you can see through the entire gear. GE and C-H told us directly that their engineering department does not agree with rating the main breaker and the rest of the gear with two different labels and that the line side should be considered for the whole gear. They mentioned the same thing about the plasma effect and arc jumping over (line/load or load/line). I do want to agree with you professionaly (I do personally), but I'm not sure I (we) understand the GE or C-H gear better than the manufacturer. Can you tell me when can I find an industry agreed understanding of this issue or a code section that states what you are saying?

Thanks for responding to my post, AF hazard safety is obviously a moving target and I wish to keep myself informed.

Marek

I've talked with people in the Square D engineering department who also generally disregard the main breaker in an MCC or switchboard when considering which device will interrupt the arc flash. As you've stated, the reasoning is that the bus runs through the gear unbarriered and can allow an arc which starts at a feeder breaker to propogate throughout the equipment, even flashing over to the line side of the main device. It is generally accepted that this is an unlikely event, but it is possible.

Thank you Matt.

I can easily contrive a condition wherein an arcing fault that starts in one enclosure propagates to other enclosures. However, what I'm more concerned with is HOW the arc is being initiated. If I am working on equipment inside an enclosure other than the main breaker section and the equipment that I am working on has an arcing fault due to something that I am doing, how does this in turn get converted into a fault across the line side lugs of the main breaker?

If someone can describe a condition where this occurs, then the point of initiation is at the main breaker. Since the arc is at that location, analysis gets somewhat tedious because unless we use overly conservative results, we would have to calculate the incident energy considering the additional working distance from each cell back to the main breaker as if the paneling is not present. Then instead of a "one label" condition, we have one label PER cell. AND we would have to consider the incident energy potential on the bus in each cell (basically the load side fault incident energy of the main breaker) as a MINIMUM incident energy rating since a fault across the bus within the cell can still occur. Although a fault could also occur elsewhere in the cell such as on the load side of a breaker within the cell itself, this will always be less than the incident energy of the line side.

Again I do not disagree with you, but are you saying that manufacturers are wrong and I should ignore what they suggested?
If then AF happened, person gets injured and the labels do not meet the manufacturer's recommedation...hmm I have a feeling they could say I intentionally put that person in harms way.
On the other hand realistically you're right and being too conservative is probably worse as we put the electrician and the facilty in a tough spot and even voltage testing confirming the gear is off becomes impossible.

thanks

No. They have an agenda, too. They are deathly afraid of product liability law suits that can bankrupt them. They will give you the most ridiculous, most extreme examples of what might happen just to cover themselves. As an example, ask one if you can safely rack out drawout gear while energized. I cannot get a single manufacturer to say yes despite the fact that this completely wipes out any reason for going through all the extra expense and effort of buying and maintaining draw out gear when bolted gear is smaller, cheaper, and more reliabile, if you have to de-energize the equipment to remove it from the system anyway.

What you need to do is to ask clarifying questions and how/when/where this would occur and what the likelihood of it happening is so that you can make a reasonable determination as to the likelihood of it happening and whether or not this is acceptable. For instance there is a reasonable risk that I will get into an automobile accident on the way home...so should my employer make me drive a tank home so that I can avoid this happening? As I walk out to my automobile in the parking lot, there is a possible chance that I will be struck by a meteor. Should we install meteor protection in the parking lot? And if so, just how big should those protective shields be? What if a 20 mile wide meteor strikes me?



What I'm asking is to consider what is an acceptable risk in anything that we do. You are describing equipment that has a very real possibility of killing someone if an arcing fault occurs. We've been noticing where I work that with 480 VAC systems, at 2500 kVA or larger for the transformer, it is extremely difficult to reduce the incident energy to an accpetable level. That's not to say that some things can't be done that carry an acceptable level of risk, even if they carry a very real possibility that if an arc flash occurs, it will kill someone. We just have to be aware of this and design the tasks accordingly.

Regardless, you've got a very real maintenance problem that needs to be addressed. I suggest you look for some way to install a mechanism that can be remotely operated at a bare minimum.

I just got through doing factory acceptance inspection on a portable substation today. It is 10,000 kVA. Even finding fuses for something that big is a challenge. The breakers in it are all shunt trip operated and have a motor operator. They are operating from an SEL relay. You can actually remotely operate them from any distance away. Once they are tripped, they have a built in visible disconnect switch with a grounding bar. All that remains is testing for absence of voltage which is clearly going to be nothing but a procedural exercise. The extra cost for these safety features over the "basic" version was less than 1% of the cost.