I have done a decent amount of arc flash work on panels and on switchgear. However, I am curious about equipment in substation yards. let's say, 15kV and above. Is it common for arc flash studies to be done on these types of equipment- hookswitches, breakers, circuit switchers, etc?

I work for a utility and we are going through and completing a study on arc flash for the switchgear and the question has been posed about what else should be done. So, what else is common practice in the utility world?

NFPA 70E pretty much stops strictly at 15kV. So is there something in the NESC that gives direction for this? Sorry, I am not familiar with the NESC portion as much as the NFPA.

Any help, comments, suggestions and ideas would be greatly appreciated.

First, I don't recommend using NFPA 70E for utilities at least as far as arc flash hazard evaluation goes for two very good reasons. First in the section on OSHA 1910 on this same forum there is a link to a recently published OSHA notice of final rulemaking intent for a much revised 1910.269 Subpart V rule which is a total game changer for utilities. OSHA specifically discusses the best modeling methods and makes it very clear that IEEE 1584 has its place but for medium voltage utility applications they strongly recommend ArcPro modeling software.

NESC is based on ArcPro and also has a low voltage table which is based on actual testing of typical utility equipment. Most of the test work was done under EPRI and seems to be very comrehensive from the reports I've read (only the publicly available ones). The higher voltage tables are based on ArcPro but as usual, if you put in your specific conditions you will get lower arc flash values in the ArcPro software.

Currently the tables in 70E go up to 38 kV. Since this is not based on any scientific data, they are being trimmed back to the IEEE 1584 limits (15 kV maximum) in the 2015 edition of 70E so you will be forced to go to another source (NESC, ArcPro, etc.) in that edition. Everything else in 70E is based on IEEE 1584 for the most part.

There is another area where IEEE 1584 does not do so good. That's in the low voltage (<250 V) category. The standard itself says that only a single test at 208 V is in the databsae of around 300 arc flash tests. So it is suspect and more data is coming out over time that continues to cast a shadow on the <250 V testing. In that regard my recommendation is that you don't try to use IEEE 1584 for that case either. Use some sort of consensus safety standard such as NESC.

Where you probably should be using IEEE 1584 (and 70E for that matter) is on 250-5 kV switchgear, panelboards, MVC's, and MCC's. Industrial power distribution is the target for this standard. Where utility distribution looks like a "70E" system, use that standard. Otherwise I would default to NESC.

For the record I work at an industrial site. However we have about 70 miles of 23 kV distribution lines and associated substations, as well as a 50 MW waste heat recovery generator. So at least a portion of our system is utility in nature and thus we follow both sets of rules as appropriate. OSHA has a letter of interpretation making it clear that this is what they intend.

Maybe I misunderstand, but are the equations in the NFPA70E not from IEEE1584? Specifically D.7? This is what I (and SKM) use for calculations for our switchgear for faults 15kV and below. Table D.1 in the NFPA70E states that the limits are up to 15kV. I didn't mean to imply I am just using tables. I am actually doing the calculations. This has all been metal clad, 5kV and 15kV switchgear.

I've found it so very difficult to actually find out the type of equipment that needs to be modeled in a substation type setting. It's not clear at all.

This is a case of do not believe everything that you read. For one thing, you are reading the Annexes and treating it as part of the Code when it's not. Second, you are using a software program that is dependent on you as an operator to understand the inputs and what it is doing. SKM is simple "garbage in, garbage out". For instance, SKM will certainly give you a calculated arc flash value for a 35 kV circuit, and it will be totally and utterly wrong. The only good news is that above 15 kV, it will put a little warning note on the output to tell you that.

IEEE 1584 claims to be valid from 208 V to 15 kV and from 700 A to 106 kA. In 70E editions from 2004 and earlier, the equations in the annex were not the same as IEEE 1584. They have been correct ever since.

However there are problems with this. For instance, in "Effect of Insulating Barriers in Arc Flash Testing", IEEE Trans. IAS Vol. 44, #5, September/October 2008, self-sustaining arcs at 208 V were created with incident energies of well over 1.2 cal/cm^2. Subsequent papers actually sometimes show cases where IEEE 1584 incident energy calculations are exceeded. The problem here is that arcs aren't stable at that voltage and tend to do all kinds of strange and unpredictable behaviors. IEEE 1584's test data set only includes a SINGLE value at 208 V which was the only time where they were able to get a stable arc to form at 208 V. This is hardly encouraging. Further, the equation widely recognized as the "IEEE 1584 equation" (the actual standard includes about a dozen models) is NOT a theoretical equation at all...it's just an empircal curve fit. So if you get outside of the curve fitted data, it gives bogus results. How bogus? At the "edges" it actually predicts a higher arcing current than the available short circuit current! This paper and several others are freely available from Mersen's web site where they have all been gathered together. Thats the reason that I have started to recognize that IEEE 1584 should NOT be trusted.

The problem is that if IEEE 1584 can't be used, then what? My recommendation is to use actual test data wherever it is available. Fortunately, EPRI and others have done a lot of work in this area and have published the results. I would strongly recommend that for the under 250 V category, use the tables in the NESC which are based on the same test data. That's also a big reason that DELETING the reference to the tables from the NESC in the upcoming draft 70E-2015 edition is a huge mistake. It leaves anyone outside of the range of 250 V to 15 kV out in the cold. Similarly above 15 kV, OSHA just recently published a very long discussion about validity of various models and clearly makes a case for using ArcPro above 15 kV but not as the voltage gets much below about 10 kV. If you use this approach it's not a calculation. In fact it's just one number, 4 cal/cm^2. If anything you may want to consider the "one transformer <125 kVA, <=208 V" rule as well which allows for <=1.2 cal/cm^2 for these conditions. However this "rule" is under serious question and the above IEEE paper is one of the papers suggesting that this rule should be lowered to something more like "<=4.5 kA, <250 V".