I am a Engineer for a large manufacturing company that has over 80 large manufacturing facilities and distribution centers. About 4 years ago we began performing arc flash studies and now are about 95% complete. With so many locations, we had a variety of companies perform the studies such as GE, Square D and other smaller electrical engineering firms, all of which had a PE stamp.

As part of these studies a short circuit study was first performed. This part of the study revealed to us that a very large percentage of our switch board and switch gear breakers are significantly under rated from an AIC standpoint. Some of the firms performing the studies noted the low AIC rating and recommended replacement, but then proceeded with the study and subsequent catagory ratings, arc flash boundary calculations and labels, while others refused to provide a calculation and label until such time that those breakers with inadequate AIC rating were replaced. The stated reason for stoping the study at these points was because that they could not model a system where they could not predict how the breaker would react at currents significantly higher than the rated AIC because the time-current curve supplied by the breaker OEM was not valid past the breakers stated AIC rating. While I agree with this line of reasoning, I have been unable to find any real data that gives us any kind of indication on what actually happens in the case that a breaker sees current above the AIC rating.

Certainly there are several possibilites, it may open normally, it could explode or it could fuse and not open. I can honestly say that we are a company that believes in doing what it is right for our employees, so we plan on replacing all the breakers that have insufficient AIC, however accross this many locations it is a logistical impossibility to do them all at the same time not to mention we are talking about millons of dollars.

Sorry for the long explaination, but my question is, does anyone have any advice on how we might go about prioritizing the order in which we replace the breakers? We have breakers 10% overdutied all the way up to over 300% For example it would be beneficial if there were some testing data to suggest that breakers in the 0-50% overdutied range have been tested and open normally and were still functional. Breakers 50-100% overdutied were tested and opened but were no longer functional afterwords. Breakers 100%
or more exploded into pieces, etc. It would be easy enough to simply prioritize by %overdutied, however I am not sure this is the best anwser when you factor in switchboard location, frequency of operation, etc. Also, we have no good reason to ever do energized work on our switch boards so doesn't that significantly reduce the risk of personell injury if a bolted fault were to occur?

Flash,

There are many different solutions to your problems but let me address some stuff first. I would look again at what they used for available fault currents from the utility, if they started with infinite values and used the main transformer nameplate data all of the calculated S/C values would be high, sometimes changing plant conditions can cause your S/C values to exceed your AIC's but to have a "majority" makes me think something is amiss here. Keep in mind (AS I am sure you have considered) that the OEM that is doing these and telling you you need to replace your switchgear is also the guy who will sell you the new switchgear, I am not saying they are doing this but you have to wonder if that is a possibility.

I am not aware of any published data on what % abouve AIC what may happen, all the OEM is saying is that the equipment will interupt faults up to the rated AIC, no promises above that, it may or may not as you said, it may have never been tested above the rated AIC. I have seen hundreds of failed breakers from faults estimated to exceed the AIC's, it is not pretty.

Now your solutions, one obvious one is to install limiting reactors, thsi may work in some areas, may not in others.

As far as breaker replacements go, some types are easy, direct replacements are available with higher interuption ratings, you can exchange your old ones and upgrade easy. Some types you will need to replace the gear. Your best solution in most cases would be to replace them with vacuum retrofits that have higher AIC's, you can use your same frames for this upgrade.

I would prioritize by doing the easiest and less expensive ones first, get more bang (Pardon the pun) for your buck, then do the others as bugets permit.

A few more comments:

While I would probably stop short of recommending replacements at 80% rated AIC, I have always adopted that "subjective" number as the max I like to see on a breaker, especially if it is prone to lots of operations as it would be on an overhead distribution system. Would consider more frequent mainrenance schedules! We do need to consider wear and tear at elevated interrupting levels. If the mfg. says 20 KAIC rated, is 19,999 ok?
I would be nervous! We know thay can fail at less than rated AIC for other reasons. Any equipment operated to its max rating is generally going to have a shortened life.

I would hope that no engineer would assume infinite bus for the study and if they did, they cannot recommend replacements on that info alone!

I would probably do the following if I were to encounter this:
1. Notify the customer immediately verbally and in writing.
2. Complete the study, if the customer wanted it done.
3. Add big notes to the final report documenting the findings.

Zog is correct in that these units can and often do fail violently, even when not stressed to the max.....or more.

I would hope so too Alan but having a majority of S/C levels above the AIC's sounds fishy dont you think? How often do you see that happen?

It absolutely deserves a second look and the assurance that the engineer did not use infinite bus. As the owner, I would require justification prior to the expense If they are making the recommendation based on that, then I would not have a warm fuzzy feeling about any of it!

But on the other hand, I have seen facilities, as have you that have had no upgrades for years or brought in old equipment from another location that was antiquated when installed. Actually could have been inadequate when installed if used.

Hence low AIC's while the power system has evolved in the last 40 or 50 years and is much "stiffer" today.

Alan

One other point is that while the conversation is related to breakers AIC rating, What is the short circuit rating of the switchgear/MCC/panelboard that the breakers are installed in?

Thanks for the replys, they are helpful.

On the subject of infinite bus being used, all of our analysis have been done using this method. With many locations accross the country, we are dealing with a handful of utilites only one of who, reluctantly gave us their etimate of what the SC current would be. The others refused either because they don't know and won't take the time to figure it out or because they don't want to take on any additional liability by taking away this "pad".

acobb, you are right about the outdated equipment and the low aic ratings. Our industry is fairly low tech and also not particularly hazardous, so mostly out of ignorance we had not paid much attention to AIC ratings until the changes in 70E forced us to, we just went with the cheaper breaker, which of course had a lower AIC rating.

wbd you are also correct, we have some SBs that are underated as well although this is not as predominate as the breakers.

As I have stated in other posts, the utility DOES know how to calculate the available fault duty, and before I replaced equipment for possibly no reason, I would press the utility for the correct info. I expect you have not pressed the right button yet.

If that failed, I would contact the local utility commission and ask for help.

It is my opinion that it would be totally irresponsible for a utility to adopt this approach when you need the actual info for your studies and equipment replacement decisions.

I would not start replacing breakers under these conditions. When you redo the calculations based on a finite bus, you will probably have a reduction of 25% to 50% in fault current. Some red flags should have went off when you were told they were using infinite bus. For one thing, the worse case for arc flash is not the 'high' fault current. High fault currents usually mean fast clearing times and low IE values. Its the low fault currents that have you input .8 secs clearing time and really ratching up the IE values.

Breakers are often interchangable in dimensions with only the model number changing. So changing the breakers won't be difficult. If it is necessary.

The other thing, and maybe someone else who knows this better can comment, but breakers are rated AIC RMS, but they are done so at a specific power factor, .15 to .2 if I remember. With most fault current calcs you are only looking at RMS, but you also need to check Assymetrical if your power factor is outside of the test rated pf. Typically, it won't be, but I mention it just so you have all the info if you are going back.

I would recommend that you do your calcs again. This time make reasonable estimates for utility MVA. You can do this with some degree of accurracy by knowing how far away from the substation you are, approximate size of the feeder, look at some of the equipment on the poles like reclosers for AIC (a good deal of utility gear is rated for 8K AIC) and at the very least I would take 15% off a infinite bus, and that would be on the generous fault side as most small plants I see 30 to 40% off with a finite bus.

Rather that replacing all those breakers, I would consider reactors, or installing current limiting fuses in critical locations. Its been a while, and I never use it, but there is a 'series' rating method but you have to check with the manufacturers to see if the breakers were ever series rated. I never use this method as I always full rate my gear, but if I remember it is still permitted in NEC under professional supervision.

When you get this straightened out, train your electrician to understand NEC 110 and the need to know AIC before installing breakers. Also, include in the panel label the following phrase, "Minimum AIC: 25K" or whatever value it calcs to.

Haze,

I obviously agree with you that they do not need to consider replacement based on the existing info....hence my previous post.

However, I do not understand your reference to power factor and assymetry. Last time I checked, ANSI standards require an RMS rated breaker to be rated to interrupt at an x/r ratio of 15 at rated RMS. But this has absolutely nothing to do with power factor.

Reasonable "estimates" of available....NOT...Do the due diligence and REQUIRE the utility to give up the info.....period....and settle for no less.

Estimates are just that and we have to do enough of that. To do the job you need the real stuff or nothing! To estimate is exactly what the utility must be wanting you to do, however I can't see why in the interest of safety.

I have also seen several of your posts where you reference "standard utility gear" at 8 kAIC. Based on my experience, does not match what I have seen in my 29 years of work in the business.

I can sum the whole thing up in one word: "LIABILITY"

I just had a debate in an IEEE 1584 meeting last month with a breaker manufacturer (one of those you mentioned). They refuse to label equipment when they find it to have an inadequate interrupting rating.

I was conducting the 1584.1 task group meeting where we are attempting to write a guide for specifying studies. A person brought this up in the middle of the discussion and the issue became very heated.

The person made it clear it is 100% a matter of liability. If something goes wrong and they knew about it, it places them in a very difficult legal situation.

All of the rest of the input from eveyone on this thread about the details of the inadequate equipment is quite good, but I wanted to address the actual root issue that you brought up.

I guess I didn't explain that properly. My point is that we get one rating on the equipment for AIC, and that KA RMS. But, we also need to be concerned that the asymetrical value of the equipment is not exceeded. We have to back calculate the asymetrical rating based upon the test procedure which specifies the power factor, and if I remember it was a range and not one value. When you do the analysis of your system, in addition to RMS you also calculate peak asymetrical to see if your system is less than the asymetrical of the equipment. You could have say 49KA RMS on your system, but equipment with 52KA AIC may not be sufficient because your peak asymetrical exceeds the calculated on the equipment. Like I said, I've never seen in any system I did where the peak asymetrical was higher in the field, then the calculated of the equpment. But for a thorough examination you have to verify.

As far as the utility MVA, what do you suggest. We all agree that using an infinite bus is the wrong thing to do as in all likelyhood IE values will be lower rather than higher. The utility refuses to provide fault MVA. So I see possible choices as:
1) File Petition with Public Utility Commission to instruct utility to provide info.

2) File lawsuit and get court order to demand utility complies.

3) Hire someone with knowledge of utility to provide estimate.

If my utility refused, I would file the petition with the PUC to compel them. If the PUC sided with the utility, I would then do the estimates. I disagree that a reasonable estimate can't be made. For one, we don't know the installation. Lets say the plant engineer walks his one 13.8KV feeder, goes two miles to the town substation, utility tells him that the wire is 1/0 aluminum, and he looks through the wire fence at the substation and reads the nameplate of the transformer that says 10MVA, 5.75% Imp. With this information I feel I could make a very reasonable estimate. Yes there are all sought of alternative scenerios and the utiltiy can change things at any time, but they do that anyway. How many utilities do you know that call every commercial and industrial account on wire to tell them they changed the substation transformer and while still 10MVA the impedance was changed from 5.75 to 4.75%.?


In spite of all the formulas and the 'scientific' approach, this is NOT an exact science because there are too many variables, and at 25% of those variables are outside the fence line and controlled by the utility. When I get the fault MVA from the utility I make an educated guess as to how permenant or subject to change it may be. If the value given is very low, I than run the numbers to see what happens to IE if it was raised by 25%. If its very high, I run it also at 25% lower, if its average, I look 25% up and down. I then use whatever value gives me the highest IE. In most cases there may be two transformers between the worker and the utility fault MVA, so a 25% change in utility fault MVA may only effect the IE at the panel by 5%. I certainly don't want to have to redo 300 labels in my plant and spend another $30K on engineering, because the utility changed out a substation transformer. As engineers, we have to anticipate what could change our values and make appropriate adjustments so that all our good work isn't lost with one utility change. Like I said in other posts, I don't worry if the 'real value of IE at time of calculation is 5.28 cals/cm2 and my label say 6.4 cals/cm2. I build the system to protect the individual, but also to anticipate the next expected 15 years of changes, without having to go back and redo the whole site because we made one change. The science is not in the number on the label, its that the PPE always meet or exceed the IE at the time of the fault.

This is just my opinion and method of approach. I welcome to learn from others.

Thanks Jim, no doubt it is all about liability. I would not label a piece of equipment that was exposed to faults in excess of its rating either, unless it was with a big "DO NOT OPERATE - NOT SUITABLE FOR CONTINUED SERVICE" or something of that nature.

Wonder what color that one should be? Maybe Fireball Red?

I think doing a study and notifying the owner of deficiencies is prudent, in fact a part of the process. One has to expect that there will be instances of under rated equipment on occassion.

...and its a good opinion.

I was just giving a little background on how I think the situation in this thread got started. Since I was co-chair and running this part of the 1584.1 meeting the debate was directed at me - only since I was the one standing in the front of the room.

In the end, I left it as we would not include any direction on this in 1584.1 the spec/guide because it is up to the ethics of the person / company performing the study.

So Jim, sounds like you ended up in the wrong place at the wrong time and got hammered?

Haze, just to clarify:
I might have mis-understood your part about AIC ratings. I figure you may have been meaning available fault current? Quite normal for a 15 or 25 kV distribution system to have 5 to 10 kA available at the sub bus, and certainly some have more, but the majority of the equipment we have installed over the past 20 years is rated for 20 or better kAIC at the subs.

Equipment out the line could be anywhere between 4 to 12+ kAIC depending on location and age. I guess my major concern for the estimating process is that depending on the actual customer location on the system, the fault duty could be extremely high or low. I know you are saying "average", but still it may not be close to the actual. I see actuals a few miles from a sub as low as 2,000 to 3,000 amps fault duty.

And given that the utility will make changes without notification, it is necessary to test the sensitivity of the installation to system changes. That is partly why I use my 80% rule for available/rated.

And you are absolutely correct that there are assumptions to be made, I just believe that available fault duty is the major initial piece of info to use and even an educated guess comprimises the entire study process.

And I as well still carry my "License to Learn".

infinite bus = invalid arc flash analysis



Flash17,
I would make the argument that all your arc flash studies performed using an infinite bus are invalid. I say this because the actual available short circuit will be lower, hence the arcing currents are lower. This puts the arc current on a different part of the trip curve (slower), therefore the incident energy will be higher. Using infinite bus is not conservative for arc flash.

I have only run into one utility that would not give the available fault but gave a range of 7 to 100 MVA. I had to use the lower value to be conservative. This lead to many Cat #4 AFH but as I told the client without better information from the utility, that is what it is.

I am always writing white papers to my PUC about utility tariffs and practices. I think the next one would be on the need for utilities to provide fault MVA. We have National Grid here in RI, but they are good about providing fault MVA. They give you all the fuse or recloser settings and fault mva.