This week's question comes after having this discussion brought up to me many times recenlty.

Most of us know the new OSHA 1910.269 has been out for a while and it has quite a few major changes regarding arc flash requirements.

Appendix E of OSHA 1910.269 references the software ArcPro by name. This program is widely used and is about the only game in town for arc flash analysis above 15 kV and single phase. I have also used it and it is the basis for some of the tables in OSHA and the NESC.

I'm not sure what the alternative language could have been but several have brought up their concerns regarding a product reference within Federal law - even though technically it is in an annex. Some have questioned whey didn't they also reference EasyPower, ETAP, SKM, etc. for < 15 kV? A similar question could apply for other international standards such as IEC etc.

Here is this week's question:

Should Federal Law reference a commercial product?
Yes
No

I'm sure this one may spark a bit of debate. (which is always encouraged)

Whether arcpro should or shouldn't have been listed is one issue.

I see a larger issue.

Since it can be 10 to 20 years between revisions, what if a software company or organization (EPRI?) develops another model in the future that is proven to be valid and a viable alternative. How do you justify using the latest and greatest when the OSHA Annex references only one "product" for > 15 kV.

I understand the intent of referencing arcpro but I think this one of those cases of unintentional consequences.

I think it is interesting that there should be any outcry against a commercial product (ArcPro) when the NESC used this data to create free tables.

I'm sure EPRI could create something and it would be well received. ArcPro was the only software based on an arc physics model rather than extrapolated from limited research data. While I applaud IEEE 1584, ArcPro was there when there was NOTHING and they have never pushed this. Kinectrics didn't even know it was mentioned in the standard until I told them. Certainly wasn't their idea.

Interesting no one is upset that not-for-profits compete with for profits and pay no taxes. ArcPro has probably not broken even on the original research. It is a very inexpensive option with lots of support from the company and they have historically upgraded it for little cost.

I'm glad OSHA did the research to show that IEEE 1584 should not be used on 15kV and above instead of what everyone had done up until this point, which was just extrapolate IEEE 1584 into infinity.

Hugh Hoagland
e-Hazard.com

I disagree partly here. What IEEE 1584 doees is present at least three models for arc flash estimates, not just a single model. This would be missed if you don't have a copy of the actual standard. The first model, a theroetical one, is the Lee model. The second model, and the only one that for some reason the standard gets credit for, is the empirical model. The third model is a group of models that are very simple in nature but attempt to estimate incident energy from current limiting fuses. IEEE 1584 itself only makes a numerical argument for the accuracy of any of the models with regards to the empirical one.

Thus to say that most practitioners are "extrapolating IEEE 1584 into infinity" and especially with reference to software packages, that's just not true. ETAP, EasyPower and SKM switch to the Lee model when it deviates outside the stated boundaries of IEEE 1584. If you simply read the results, you'd miss this but at least in SKM (the one I use), it pops up a footnote that explains this.

I'm happy for one with what OSHA did. Let's say for instance that you are an engineer at say a large mining facility with a cogen that has internal distribution systems well over 15 kV. So the company gets the "arc flash bug". Now what? Well, we could appeal to the tables in 70E for this one case. Oh wait, as of 2015 any reference above 15 kV is deleted. So then we've got to find an alternative. My arc flash software prodcues a "result" but it is nonsensical (Lee) and I know it, and the results are patently ridiculous requiring 150+ cal/cm^2 suits in the main sub just because of the voltage. So what can I use? I could use NESC at least for open wire but what about the indoor, enclosed stuff? And if I just use a commercial software program (ArcPro) what justification do I use compared to Lee other than an appeal to sanity? And even failing all that, what if the worst happens and someone gets hurt and the company gets sued, and this issue comes up? I can hardly say that I'm following recommended practices from anyone, anywhere. Worse still, what if I'm an engineer at a small local REA utility and I've never even heard of this before and need to get up to speed quickly, and the local engineering firm is telling me that my guys have to dress in 100 cal suits and use 35 foot long hot sticks even to operate a recloser handle? What then? And yes, Lee really is that silly and that bad.

So, I am in agreement.

So lets say that a competitor to ArcPro comes along. Then what? Any competitor is really going to have to prove their results. That means comparisons to ArcPro, Lee, Duke heat flux, and hopefully real, measured data. Frankly, the fact that OSHA published this in an annex changes NOTHING. The same burden of proof is still required to use Brand B. This is much different from placing it in the text!

Finally as to the for-profit vs. the not-for-profit argument, what exactly is OSHA supposed to do, or anyone for that matter? Prior to the annex, we had a table of values in the NESC that descended from a mystery source that we can't talk about, or go out on your own and do the same analysis. Had OSHA put together the same table and simply labelled it as "Lee theoretical model", "Brand X" (Duke), and "Brand Y" (ArcPro), we'd be stuck at the same situation. Without know who brand X and Y are, it's just a meaningless comparison that can at best show how similar the results are and not for the purposes of software selection. ArcPro at least claims (although if you get the actual papers you will realize the information is incomplete) to be based on peer reviewed models. Duke heat flux is a lot more vague than that.

For that matter, IEEE 1584 empirical model is just that, a curve fit. There is ZERO theory behind it beyond statistical analysis. Yet we appeal to the empircal model as if it's the holy grail and reject alternatives which may be based on physics based on how well they compare to IEEE 1584, rather than to the underlying data. We have erected a substantial hurdle in the research space for even a 480 V arc flash model.

Looks like the survey results speak for itself. I bet any of the other software manufacturer's would have loved to see their name mentioned. What great P.R.

The old adage "it is not important what you know but who you know" is always true.

Where did this come from? IEEE 1584 has ALWAYS been for systems with voltages from 208 V to 15 kV and NEVER was intended to be used above 15 kV. I know of no one that has ever attempted to extrapolate IEEE 1584 kV and I work with many people including people from the major software companies that are on the IEEE 1584 Working Group

From IEEE 1584 Section 1.2
1.2 Purpose
... Applications cover an empirically derived model including enclosed equipment and open lines for voltages from 208 V to 15 kV, and a theoretically derived model applicable for any voltage.

From IEEE 1584 Section 5.4
5.4 Lee method
For cases where voltage is over 15 kV, or gap is outside the range of the model, the theoretically derived Lee method can be applied and it is included in the IEEE Std 1584-2002 Incident Energy Calculators.

IEEE Pretty clearly states that for voltages above 15 kV, the Lee method may be used (however as most of us know, this is not a good approach due to its extremely large results at higher voltages.)

From IEEE 1584 Section 9.1:
9.1 Range of model
The model is applicable for systems with:
— Voltages in the range of 208 V–15 000 V, three-phase.
— Frequencies of 50 or 60 Hz.
— Bolted fault current in the range of 700 A–106 000 A.
— Grounding of all types and ungrounded.
— Equipment enclosures of commonly available sizes.
— Gaps between conductors of 13 mm–152 mm.
— Faults involving three phases.

The associated tables in IEEE 1584 for gaps, equipment and distance exponents don't even go past 15 kV. In addition, the major software companies switch to the Lee method above 15 kV. The voltage range for the applicability of IEEE 1584 has been published and widely known and adhered to since it was first published in 2002.

Well, you can use the summary version in the 70E Annex or modify the spreadsheet to remove the limitations to do it. As you said, none of the major software vendors do this.

Two caveats to doing this:
1. Its a curve fit. Towards the edges of the curve fit, strange mathematical errors appear such as predicting arcing currents that exceed bolted fault currents.
2. At the low end in particular it tends to overpredict since there was only a single data point at 208 VAC and arc stability becomes a major factor.

Just because you can put numbers into a formula and get results doesn't mean it's right, even if you are doing it within the defined limits. Moving out side those limits is even worse and can't be called "IEEE 1584 standard" anymore.

I voted for It Depends for several reasons. One is that in areas where I have done work before I already have established contacts and a relationship with the engineering staff, so it is easier. I have found that the larger the utility the more difficult it can be as not all utilities have an Account Representative assigned to industrial accounts. With larger utilities it takes a while to get to the engineering staff and the first response is from Customer Service where the maximum possible short circuit current based on transformer size is provided.