We have a mcc with a main breaker labeled at a 4 up to the line side of the main breaker. Does this mean all the mcc is rated at a 4? Can we troubleshoot in this panel if we are only qaulified to a 2 for work ?

I'm assuming you mean "H/RC" rather than cal/cm^2.

It's a little debatable. Let's suppose for a moment that you are working on something downstream of a breaker. What is the likelihood that you will accidentally trigger a failure upstream of the breaker? What is the chance that the upstream section will have an arcing fault of it's own that just happens to coincide with the work going on? And if it does, is your location inside or outside of the arc flash barrier in the location where the proposed failure is going to be happening?

Your answers, which are clearly dependent on the task being performed, change the result. Note that although structure (panels, doors, etc.) clearly does have some effect in so much as they may perform as barriers, there isn't a lot of scientific data on their use as a heat shield and/or ability to deflect some of the incident energy yet.

The last part of the comment leads to very good Question. An accident occurred a while ago near Reading PA. It was described at a meeting of IAEI in Central Pa. Two workers were in an electrical room prior to beginning work to replace a circuit board. The voltage in the compartment they were to work was less then 50 volts. However prior to opening any thing a bang occurred in the 480volt equipment behind them, followed by a second bang and an explosion. I believe one person died. At the time the accident was described, the second person was off work for a year; he was just riding around with a colleague to get out of the house, the hot fumes from the explosion behind him had burnt the mucus membranes from his nose and throat completely off.

Tests by the navy have determined that almost a ton of force can occur behind the barriers in equipment. The old heavy metal doors in older switchgear, might, or might not contain this amount of force, I don't believe the light sheet metal covers on some of the newer equipment will.

Without question a tragic accident, however; no amount of electrical safe work policy and prevention will prevent a random accident from occurring in nearby equipment. The accident you describe is functionally similar to kids leaning their bikes up against pad mounted transformers serving a public library.

I agree with Paul's answer that the safety depends on the work type being performed and construction of the equipment.

Obviously this is an opinion, but, it is backed up by over twenty arc flash accident investigations. You have several issues:

1. The incident energy values that most programs generate are based upon an arc fault on a vertical bus end. In a typical arc flash the arc runs away from the source and angles down due to the magnetic forces. Find some video's of the IEEE arc-in-box testing. Many MCC faults do not follow that pattern and are defined as a horizontal bus fault, there have several IEEE papers (Mike Lang/Ferraz Shawmut) written on this variation of the fault.

The arc flash is now more horizontal and the arc current is reduced, therefore, the incident energy typically rises. In short, most engineers do not take this into account when performing calculations and the result is an incident energy calculation that is lower than actual. Easy Power has a switch that allows you to define a fault as a horizontal type or barrier type of fault. I am not sure if other programs have this ability.

2. In my opinion you can never use the main device in a MCC to lower the incident energy. Once the rear compartment fills with the conductive copper plasma the fault propagates to the line side of the main overcurrent device and it will not trip. My assumption is very easy to confirm, ask any MCC manufacturer to put in writing that their specific main overcurrent device will lower the incident energy within the MCC - they will not respond. To further strengthen that position, I have investigated a death that occurred when working on an MCC . In that case the calculated incident enrergy did not appear to model correctly until the horizontal bus modifier was used.

Taking item #2 to the next step, the same concept applies to switchgear, the main device does not lower the incident energy, unless you specify rear connected switchgear type construction.

Hi Flash,

Can you expand on this part of your comment?

Thanks

[

It actually says a couple things. That bus configuration matters at certain distances but as you get farther away, it all tends to produce the same result. So configuration matters SOMETIMES.



I work in the mining business. It's an unfortunate fact of life that mining companies tend to "worst case test" their equipment a LOT. I've seen plenty of cases where the air heated up to the point where the CFO voltage drops down and you progress from L-G or L-L into LLG or LLL faults. But I've never seen any evidence of plasma vapor or gas that causes this. From a physics point of view it's not strictly necessary for this to be the case because the conductivity and more importantly, critical flashover voltage, decrease dramatically with temperature increasing. Copper oxide is a semiconductor. Though lots of maintenance departments try to reuse equipment that has been burned badly, if it was all sprayed with Cu (which rapidly turns into CuO), you could never get it to pass a megger test again. I think Hugh Hoagland also mentioned something similar in his testing.



Schneider, Siemens, and ABB among others specifically publishes white papers describing this approach to lowering arc flash values and so do the relay companies such as Schweitzer. Your comment doesn't seem to hold up.



I'm not entirely sure how you could have known what the incident energy was. IEEE 1584 specifically states that if you follow the standard and use PPE where the cloth (not necessarily the clothing) has been tested with the ASTM 1959 standard, that there is a 90% confidence of a person walking away with less than a second degree burn. IEEE 1584 is very up front about the fact that the arcing current is the most troublesome value to model and that there is a long tail distribution on the arcing current so it is entirely possible to have an extended arc resulting in higher incident energy values. How would you possibly know whether or not you accidentally tripped over a "10% case"? For that matter I'm not aware of any case where the method developed in IEEE 1584 has resulted in too low a value and consequently an injury, and I'm sure that the IEEE 1584 committee would be dying to hear every last detail themselves if nothing else than to validate the "tail" part of the theoretical data that their calculation method has trouble modeling.



I'm also not quite understanding your comments here from a plasma point of view. If we take two parallel buses and place a conductive object across them and energize it...a rail gun...the magnetic force induced in the object propels it away from the current source, not towards it. In fact there has been some preliminary research done on arc flash in overhead lines and this is one of the major stumbling blocks...because the arcing fault can travel significant distances down the line to locations well outside a calculated "arc flash boundary". So far there are not a lot of good answers for this issue but one thing that is clear is that the fault travels away from the source.

I would also be very wary of doing any kind of investigation especially one that may involve potential legal concerns such as a fatality investigation by using an equation or model in a software package that is not thoroughly documented. ETAP may have a "horizontal bus flag", but unless they publish the modified equation and it is pretty well peer reviewed, I'd be very nervous about using it. That is the primary reason for sticking with either the tables in NESC or 70E, or using the calculation methods in IEEE 1584. Otherwise, you are putting yourself on the line in terms of liability even if you use some other programmer's "guess" as to the correct way to model things.

If you can show a way for a failure to propagate to the line side of the breaker such as in a 480 V panelboard then no. But in the case of an MCC although there isn't great isolation from one bucket to another, the main breaker is generally isolated so in that case, I'd model all the buckets except the main breaker bucket using the incident energy available on the secondary side of the main breaker. The main breaker of course is controlled by the next upstream device.