Question: Arc flash labels often list incident energy with one or two decimal places. How many are necessary?

Answer: From a mathematical standpoint, you can display as many decimal places as you wish but let’s avoid competing with the number of decimals are in Pi.

What is practical? It is important to remember that IEEE 1584 describes incident energy as an estimate. This is due to many factors that influence the actual exposure—such as the exact point of the arc within the equipment, the exact electrode gap at that moment which can vary as conductors melt or be blow apart as well as many other system-specific conditions.

In practice, most labels use anywhere from one to two decimal places. It is important to keep in mind that adding additional digits gives a false impression of increased accuracy, even though the underlying calculation remains an approximation

This may be blasphemy and self-defeating since most of what I do for a living are power studies, driven by arc flash studies and the need for labeling, but I think it is really a rough estimate at best in many cases (see my video on industrial control panels and my video on pad mounted transformers for two examples.). There are just too many variables to be able to accurately calculate IE at the working distance for every installation ever conceived. I don't blame the IEEE. It is obvious a lot of work has been put into testing and developing the formulas. What I have seen of the test setups though, they are necessarily generic. There is no way to duplicate every type of equipment with every type of method to initiate a fault with every range of available fault currents and every point within a enclosure and every bus gap etc. It is impossible.

I have also been privy to dozens of studies performed by dozens of other engineers and all of them had problems to say the least. Not saying mine are perfect, or done "the right way" every time without mistakes. I am just pointing out there is a lot of potential for errors when doing these studies. There needs to be a better way.

I think the industry should push more towards making safer equipment and affordable retrofits for mitigating risks to increase safety as well as stricter codes. Of course, I know that cost is a major factor there that cannot be ignored. I know for instance, there is fully insulated MV gear (PremSet) that looks to me to have a very low potential for shock and arc flash hazards, at least compared to uninsulated gear. That technology doesn't seem to be widely adopted or in use though.

Not really sure what the answer is. Would like to hear others opinions.

Thanks for your thoughts! Opinions are usually all over the place.

I tend to favor one decimal place. PPE tends to have zero decimals but if the calculation uses zero places, was the calculation rounded up or rounded down? Also 1.2 cal/cm^2 is often cited by standards and has one decimal place so do what do you do if you round 1.45 down to 1.0.
Quite a bit to consider.

My choice would be 0 decimal places with all numbers rounded up to the next whole number.
I am amazed at the people that will want 39.99 shown on the label because it is less than 40.

Other than the old 1.2 cal rating nothing uses any decimals so do we need to keep doing thing the way we did before we knew better.
Training is supposed to be updated regularly, so doing the things the way grandpa taught us doesn't continue to make sense.

This forum needs a "Like" button.

This reply serves as a "like" to bbaumers reply.

I voted for one decimal place. My reasoning was the same as what Jim shared. This allows a QEW to understand the calculated IE best and not question how rounding was done. As long as we identify 1.2 calories as the threshold for a second degree burn and the point at which arc-rated PPE is necessary I do not see how we can not identify the IE without, at a minimum, one decimal place. Going a step further, I do not see great value in expanding to a second decimal place especially given the amount of variables that go into the calculated value.

Everyone involved in the verification of data, modeling, analysis, etc. should be properly trained and qualified for what they are doing. With that said, it can be impossible at times to be exact with conductor lengths. Utilities do not make fault current contributions easy. Many will not provide any information. Many that provide information will give secondary fault current values derived from infinite bus without indicating this. At times equipment characteristics (transformer impedance for example) cannot be confirmed. Assumptions are made at different points which lead me to see little value in listing more than one decimal place on an electrical hazard label.

Maybe it is easiest to follow the PPE standards - this is taken from ASTM F1506 (it is also included in similar fashion in ASTM F1959):

As my professors always said "THREE SIG FIGS!! That's the max you need/should use for calculations involving estimation, and even that might be overkill in this situation. That 3rd significant figure is only going to affect the result by 1% in marginal cases. You will never need 2 decimal places after the period.