Happy New Year!

The first question of the year is about the level of detail for arc flash calculations - it is a bit hypothetical.

The amount of incident energy that a worker receives from an arc flash can be influenced by the size of the enclosure and location of the arc flash within the enclosure. If the enclosure size was factored in, it would require determining the enclosure/cubicle dimensions for each case. The 2002 Edition of IEEE 1584 is based on default enclosure sizes that were used during the testing. If there was a method to factor the enclosure size into the calculations, would you use it?

If there was a method to factor the enclosure size into the incident energy calculations, would you use it?
Yes - I'm up for determining sizes
No - I would prefer a default value
Maybe/It Depends

...and of course, comments and discussions are strongly encouraged!

Although one may consider more detail is better, I believe determining equipment dimensions, especially if drawings are not available and measurements are required, would be very impractical.

I can see the time and cost of a study skyrocketing if you measure everything. What about equipment with different size sections/cubicles like switchboards and motor control centers? Would you model each different size section or cubicle individually or use the worst case?

What happens is that the amount of "focusing" of the thermal energy changes dramatically depending on enclosure shape. The current table reflects essentially the cases of open air, "small enclosures", "medium enclosures", and "large enclosures". It does not consider vertical vs. horizontal electrodes nor "box/barrier" configurations which can increase the calculated incident energy by a large factor (200%+).

Panelboards are relatively well covered because they are large enclosures but relatively shallow so that the incident energy is affected by only a small amount relative to the open air case. It would be incorrect to use the dimensions of a "compartment" in panelboards/switchboards since they are wide open internally in reality.

MCC's are not well covered because the model is based on roughly a 20"x20"x20" enclosure which is representative of "1 high" standard MCC buckets or a disconnect or combination starter across the board but since the most significantly arc flash hazard in an MCC would occur on the bus at the back of a section the correct model would be the dimensions of an entire section which is typically about 96" tall by 20" wide. A few are made which are only 16" or 24" wide and some are reduced height so they are only 84" tall, and some are reduced depth (16") but these are relatively rare and the dimensions are not that drastically different. The net result is that the current table entry for MCC's is probably wrong to begin with because the major arc hazard threat isn't a failure inside the bucket itself but back behind the buckets.

And when we get to medium voltage and switchgear, there are few standards for dimensions of any sort. Depths can vary anywhere from about 36" to 60" for "standard" gear with widths from 24" (for GIS...unlikely to be analyzed) to as wide as around 60" with the most common widths of 36" or 48". Heights are usually either 96" or 48" (for double stacked gear). Arc gaps are so widely varying that for practical purposes it should be measured but is usually roughly similar for a given voltage class.

So at least with panelboards we can still use a table approach. With MCC's unless the equipment is somewhat specialized (VFD sections are typically 24" and everything else is 20" wide), the width and height rarely changes for a section. It is probably a waste of time and money to model individual buckets because the incident energy from a bus bar arcing fault is greater and present for every bucket in a section. Similarly across multiple manufacturers and types of equipment there is nearly a universal standard of spacing terminals with a 1" gap for 480/600 V equipment. It sometimes varies for 208/240 V equipment. Above 480600 V there is a lot more variation overall but even though there is no set standard for the most part equipment tends to be similar in terms of arc gaps for a given voltage class. For instance 5 kV rated equipment used for 4160 V and 2400 V tends to have terminals with about 3.5" of open air space between them. This dimension seems to hold true for Cooper, Linepower, Eaton, Allen Bradley, Toshiba, Siemens, and Powell gear that I've checked. Similar nearly "universal" dimensions seem to exist for the 8 kV and 15 kV classes. The 28 kV class seems to vary a lot and again it's pretty standardized for 35 kV class. It holds true for things like lug spacing for circuit breakers and disconnects. But it seems to vary all over the place for transformer lugs and busbar arrangements. And applying the proper arc/bus gap dramatically changes the incident energy.

So I can only vote for one option but I think that we need both solutions. We need a table of values and this is appropriate for a lot of equipment. But the calculation approach should be available when it is more appropriate or where there is a desire to model particular equipment (especially custom equipment) more closely.

No - I would prefer a conservative default value. I'm guessing the results don't vary too widely, and a conservative default size might be chosen.

I'm all for more accuracy, however, I fear this could turn into a can of worms. This issue is partly why I use a safety factor in applying the results of an analysis.

Over and above the built in safety factors?

Have any of you done calculations for different size enclosures? If so, how about posting the data so we can see the actual impact(s). Then, we'd have a real sense of the effects.

I understand the perception of more detail is better and I'm sure the size affects the incident energy, but wouldn't this become quite complicated and expensive?

OK let's be clear about what we're talking about. Although the more recent testing may expand this the original IEEE 1584 testing was based on 3 enclosure sizes:

12x14x7.5
20x20x20
45x30x30

The IEEE 1584 proposal may be slightly different but the public one that includes details on a box is R. Wilkins, "Simple Improved Equations for Arc Flash Hazard Analysis", IEEE ESW-2004.

In Wilkins calculation first calculate arc flash for an open air condition. The apply the following correction factor:
E_box=k*E_sphere/(a^2+D^2) where D is the working distance and a and k are constants that are derived for the different enclosures.

Fontaine has a variation of what is essentially the same formula where the form is:
E_box/E_sphere = k*r*pi/(1+(a/D)^2) although the value for a is given as A=a*another constant

I believe I've seen something published along these lines that gives a means of calculating the constants k and a based on the inside area of the enclosure or some similar known physical parameter that is fairly easy to determine. The resulting formula is no more or less complicated than the original exponent and multipliers.