Hi,
What is the correct procedure for posting AF labels? I believe that Incident Energy is calculated for the panels/bus. and not for the feeders.
But still some folks are saying that the AF labels should be posted for each feeders connected to that panel. Is it correct?

The incident energy is for the entire panel/switchgear/MCC that is a result of contributions from each incoming source of fault current which would include motors at the end of the feeders.

So there is usually one worst case incident energy that is found and the panel/seitchgear/MCC is labeled with that value. The individual feeders are not labeled as the IE would be the same as the whole panel. For example: In an MCC a bucket that is feeding a panel 100 ft away, has the same IE as the MCC as a whole. So what is the benefit of labeling each MCC bucket?

As long as it's clear that's what matters. ANSI Z535 is the standard for safety labels in terms of layout for instance but doesn't give specific locations for it.

Agreed with the previous mention with one caveat. There are two areas where things go a little sideways on you. The first one is when it comes to combined labels with both arc flash and shock information on them. When you have an enclosure with multiple voltages present and the most obvious case is with transformers where there is only a single compartment such as a lot of dry transformers, you will have two voltages and two incident energy values. Logic says to take the highest one of each. The software generated labels will of course put the higher incident energy on the lower voltage label and vice versa. So you either have to apply both labels together or create a custom one to fix the error.

The second problematic area is with gear that includes a "main". With relatively open equipment such as a panelboard it's pretty easy to figure out how to label it because as stated the incident energy and voltage is the same for the entire panel. But with some types of gear, especially switchgear, where a "main" breaker or fused switch is isolated from the remaining lineup you will frequently have two different ratings, one for the main (or multiple "mains" in the case of generators, double ended gear, rings, network cabinets, etc.), and one for "everything else". Depending on the layout this may again be pretty obvious but with some gear especially switchgear with lots of compartments, mains, etc., the safest approach is just to use one label for every compartment.

The third issue with regards to placement is when you have multiple ratings for different conditions. This would occur for instance if you label equipment with one rating for one power source and a different rating for a backup power source, or if you have double ended gear and the standard operating mode is to operate single ended but you want to label it with provisions for having the tie and both mains closed (most severe incident energy rating), or some other atypical combination that comes up frequently enough that it needs to be labelled for this condition. In this case you may have more than one label as a simplified approach to simply having one big label with a "table" showing different conditions. Choose your poison.

In practice I simply print one label for every bus for both "feed" and "load" sides. Then I simply toss the labels that I don't need. This increases printing costs a biit but the time involved in trying to get the number of labels "just right" from the office is more time consuming and costly than simply figuring it out in the field.

I am not sure if I correctly understand your question but I have seen where large MCC's have a label for the MCC and then a label on each feeder bucket. I do not think having a label on each bucket is necessary, especially if there are no "maintainable" loads such as a motor without a disconnect. If there is such a device then the label should be on that device not the feeder breaker.

Too many labels can be a sensory overload and often means people eventually stop reading them.

In my experience, when there are labels on every MCC bucket or Busway plug, I consider the electrical safety program, per MFPA70E, to be suspect.

Really? Have you read NEC, the 2014 edition. In order to comply with the new requirements especially with mains in double ended switchgear you will need:
1. Shock hazard label. Multiple labels if there are multiple voltages present especially if there are different compartments, or if a capacitive trip device is in use which is becoming even more common as newer switchgear switches over to using magnetic actuators.
2. Arc flash hazard label. In all likelihood there are at least 3 of these. One for the normal case with the tie open, one for single ended operation, and one for double ended operation with the tie closed.
3. Voltage label(s).
4. "More than one source" label indicating the tie/second main.
5. "More than one source" label for the DC power supply.
6. For medium voltage, yet another "more than one source" sign if it's a fused disconnect for the fuses themselves.
7. Ground disconnecting warning label.
8. Various interrupting rating and other equipment rating labels.
9. A shock hazard warning label on the back side or other compartments.
10. A "high voltage" warning label on the door to the switchgear room even if all the equipment is enclosed (stupid but that's the rule).
11. An inspection/date label. Not required by NEC but it's almost impossible to manage compliance with NFPA 70B/NETA MTS (arising out of Chapter 3 of NFPA 70E) without it.
12. Identification of the purpose of the equipment, labels for all the lights and controls, etc., as well as an overall enclosure label. This is actually an NFPA 79 and OSHA requirement but is so pervasive that it should be mentioned. Although this entails lots of labels I'll be nice and lump it only into one category.

That's all the basics. And by the way, the signs need to be ANSI Z535 compliant. This is complicated by two requirements of ANSI Z535:
1. The wording, signal word, etc., are in DIRECT CONFLICT with the specified wording in NEC. For instance "DANGER -- HIGH VOLTAGE -- KEEP OUT" comes up a lot. Here are the issues:
A. The signal word must be in Z535 compliance. In the case of the signal word DANGER it must be a situation where a fatality or serious injury is going to happen without taking other precautions. An entrance to an electrical room with enclosed gear is only supposed to have WARNING or NOTICE, not DANGER.
B. The sign must describe the specific injury and action taken to avoid it. "HIGH VOLTAGE" is not an injury. SHOCK or ELECTROCUTION is the injury.
C. ANSI Z535 specifically states that NO MORE than 3 warnings can be used per task. Granted all 12 of the warnings above do not apply to every single task but you get the picture that it has gotten completely out of control.

Responses like yours may be why there seems to no longer be 'two-way' discussions on this site.

The OP asked about feeder buckets. I responded with a comment that specifically mentioned Motor Control Center buckets and Busway plugs. You chose to change the topic, in what appears to be a post intended to called attention to yourself while belittling someone else.

No offense intended at all. Several posts recommended one label per panel. Others pointed out that this gets ridiculous very quickly. And it is typical on forums like this to get a diversity of opinion. Mine was that the current NEC standard for label requirements is vast overkill even without putting on one label per panel, doesn't comply with the very standards that it references, and likely will be pared down in the future. Still...

Depending on the arc flash study, you can have between about one and four incident energy ratings groups. There is no specific rule on how to handle this. If there is a "main breaker" or some other main overcurrent protective device the incident energy for that section (or at least bucket) is determined by the line side. If there is more than one feeder, there may be an additional incident energy rating. The rest of the MCC is rated based on the possibility of an arc flash on the bus bars and carries a single incident energy rating, unless again the arc flash study somehow breaks it up into more sections. Finally the external devices may have a separate rating. The assumption made by many is either than in general the incident energy is low enough not to worry about it (an assumption that is usually but not always correct) or (logic error) that the incident energy is always lower than the incident energy of the MCC. Thus the final label for an incident rating is not for the MCC itself but for the field device. With many field devices (e.g. motors) that are frequently replaced, some sites place a "load side" incident energy label on the MCC along with the "line side". The MCC is also normally the one and only lockout point so it makes a lot of sense to have the information handy at that location even if a test for absence of voltage is done at a separate location.

Whatever the arrangement is, the key point is that it should be part of the training and it should be clear. The standards don't give a specific placement guideline other than to specify which equipment must have a label, only rules as to contents and formats. But what I usually see is that there are one or two stickers placed right next to each other with different ratings and often some less-than-clear description of what it means. So it becomes confusing as to whether the label is for the main breaker, the MCC, the load, or it's just an old sticker that was never removed.

I'm a big proponent of the single label for the whole MCC standard, except that I like to place a second label on the main incoming section if there is one. Arc propagation CAN happen but I haven't seen any evidence that it can happen from one MCC section to another, unlike panelboards. If the arc flash study says otherwise, then I'd deviate from it. It is also nice to CHECK the rating at the downstream equipment on particularly large loads or long cables where the incident energy can be much higher than expected, but clutters up everything and makes it confusing if it is not absolutely necessary to add labels for the loads. Since it significantly drives up the cost of incident energy studies and rarely produces significant results few of them actually bother to include incident energy studies at the loads anyway.