Doubts about the future of Arc Flash Analysis

Getting utility information has always been tough, but some utilities are adopting a policy of providing generic information only. Recently, our company spoke with the engineering standards group for a major utility company out in California. The position the utility took is that since fault current & impendance is dynamic (with switching, etc.), the utility company may be liable for giving this info if the system changes and an event occurs. So essentially, the utility company does not want to provide info that's supposed to increase safety. I'm not aware of any regulations that compel the utilites to provide the information for arc flash studies (and was told there were not). Asked how we should use the generic info given: "Find a value between 4.5 MVA & 100 MVA that gives the worst-case result."

Is this really how an arc flash analysis should be done?

Obtaining Utility Data

Some utilities are great, some not so great. A lot of them are taking the easy road - but in my experience it has been a lottery, whether you get to the right engineer or not.

I agree with John. Some utilities are very helpful. Some are not. To me it seems that the larger the utility is, the less helpful the information they give is and the longer it takes to get it.

Sometimes its slow to come but nagging I have always got something. On the other hand, who knows how accurate the data is.

The real question is should we use the information the utility is willing to give us? I mean this as a general question. Utilities are constantly upgrading and changing their infrastructure. Unless you are a steel mill or an aluminum smelter the utility will never notify you of what they are doing and how it effects your utility contribution. I wonder if it is wise or even prudent to take the utility MVA as gospel when they give it. You would need to be requesting the info at least annually, and that would probably tick them off. The utilities are not going to take on the role of notifying every customer when they make a change at a substation. Would not a better approach be to run the analysis with an 'infinite bus' and some absolute low value for the 'finite bus' (like maybe a 30% reduction). I don't know what the value would be on the low side because I don't know the industry. But if we could agree on some value that would be an absolute 'low' and we know the 'high' -then take the worse case - wouldn't that then cover any and all changes in-between.

I disagree with the utilities refusal to give out the proper info and feel they should be compelled to do so. I also agree that the end user must realize that the system is dynamic and should model the facility for additional cases to determine how sensitive their equipment is to utility changes.

We always advise our clients to furnish the info.....with the caveat that the info is subject to change without notice and should be used accordingly. I would rather start with the actual/known figures and then make the assumptions for changes.

In many cases when a customer is sizing service equipment we will give them infinite source, and tell them so, but for arc flash work we give them all of the numbers.

I agree with acobb: The utilities should be compelled to provide specific information with a corresponding disclaimer / release of liability.

Utilities will probably start charging to provide this information (some already do), but I think this helps the arc-flash industry. It gives the utilities more incentive to work with us and provide just the information we need. If a client requests updated arc-flash (as mentioned by haze10), it wouldn't tick off the utility co. if they are getting paid again.

In an ideal world, the utility would give up-to-date info and a range of past recorded info (some utilities regularly document the system state I hear).
From these ranges a realistic worst-case scenario can be built.

My biggest worry now is the liability issue brought up by utilities.

Let me ask you this, if you had to pick a worse case for lowest fault current, how would you do it. Let assume 13.8KV delivery line that is run along highway or street, what would be considered a utility local feeder. Would you use an MVA value, (say 20MVA) or use a % reduction off your infinite bus (say 40% reduction)? I'm just trying to figure out an alternate method to look at the worse case. I can make a judgement call that if the facility is located in an industrial area and fed from the street, that the utility would not be making a 'ten fold' change. So unless I see the utility building a new substation, the wires from the existing substation will be the limiting factor and any change at the substation would only have a limited effect (50% to 200%).

Can some of you utility guys give me some general input. I will of course start with the fault MVA provided, but since I want worse case, I'll run another calc. Would you make a reasonable assumption that fault current is more likely to go 'up' rather than 'down'. I would not suspect most utilities are changing out transformers for 'smaller' than existing. But there could be trends for less impedance in the xfmr for greater regulation.

All thoughts welcomed. Maybe this is more than is necessary and I should just go with one calc which is based on utility provided MVA.

From what I have been told, utilities change out transformers only when load usage increases (or they go bad). I was told they don't usually inform the customer of these changes. I highly recommend experimenting with different utility values to get a good understanding of how it affects the arc flash energy. Remember, you are really concerned about how the current gets interrupted by the main. With very high current (infinite bus), the main breaker or fuse probably will clear quicker. There is a point where a lower current will increase the trip time enough to significantly increase the arc flash hazard. Beware of this...I don't think utilities understand this and this is why they may give overestimated values.

Generally the trend over the long haul will certainly be an increase in fault duty from upgrading a transmission line, an existing sub, or building a new one closer to the facility. Also as load grows, wires get changed out to larger ones, service transformers get changed for larger ones also.

Also the utility will "normally" serve the load from the closest possible location given adequate capacity. However during periods of equipment failure or maintenance I have seen many cases where the load is switched to a source that may be say...... 5 miles away instead of 1 or 2 miles with a corresponding decrease in fault duty.

Much weaker supply for that time period and could take you from inst trips to time delayed with the lower fault duty. I think it would probably be a good idea to run a model that has fault current low enough to be below inst trip level to see the IE as well, and expect most do already.

As far as an upper limit calc, maybe 25 or 50% above the utility supplied data? Not sure about that one.

Don't know if I would worry too much about an increase until I saw the higher numbers from the utility, or saw the equipment going in at the end of the street. Then I would ask for new numbers.

I would also ask what kind of data was supplied, infinite source or actual data. If you could also get the service transformer size and %Z, a quick check will let you know what was supplied.

If you are getting fault levels at the secondary of a utility transformer, then changes to the utility primary system will have little effect on the fault level. Most of the impedance is transformer impedance. Smaller transformers do not have standard impedances, so if the utility has to replace the transformer without increasing the size, the impedance may increase.

You need to know how the utility has determined the fault level. If it is a maximum value, they may have used the smallest possible transformer impedance to allow for replacements or to avoid having to determine the transformer impedance at a particular location.

If you are getting fault levels at the primary distribution voltage, then the fault level could be decreased during planned or unplanned maintenance because of line switching. It would be difficult to estimate how much without knowing the switching possibilities for the primary circuit.

Absolutely agree with jghrist that the service transformer will be the predominant issue that will drive the fault duty. However I have seen source changes easily make a 10+% difference in the fault duty from the utility.

If you are even close to the inst trip, still think a lower utility contribution check is prudent. Good CYA.

Might even take it to 20% less utility contribution for worst case. Would have to take it case by case.

System maintenance issues can be a week to weeks, equipment failures can cause you to be switched for months.

I think this reconfirms what I am saying. But let me clarify some points. Most of the 'facilities' I am taking about are primary metered. The utility is suppling 13.8KV from overhead lines from the street. The facility owns the step down transformer and it will usually be 13.8 to 4.16KV or 13.8 to 480V going into some type of switchgear or switchboard. Main transformers are from 2MVA to 10MVA. But this doesn't really matter, as we are talking about designing around possible changes to the 'upstream primary'.

My goal is to NOT get to the point that fault current is reduced to a level below the Inst trip value (although usually do not have Inst trip on the Main just the branch breakers of the main switchgear (switchboard).

If I calc the Arc Energy with an Infinite bus, then it doesn't matter what the utility does. They can install bigger wires and transformers, but they can NEVER exceed infinite! So I have that end of 'worse case' covered.

Now what I want to do is to 'anticipate' the lowest possible finite bus that could ever occur (within good judgement). I need this value so that I can set the Inst trip value on the breakers to still react at this reduced level (assuming it doesn't cause any downstream problems). If I can get the breaker to still trip here, I've reduced IE. If I can't set the trip that low, then I still need to calc IE for this worse case whatever the trip time turns out to be.

So the question remains, how low of finite bus do you go as a rule of thumb. That rule would change depending on the installation. If I was tied to an Interstate Feeder at 369KV, I'd use the fault MVA the utility gives me minus 10%. But if I am on the street at 13.8KV, and the utility switched me over to next substation to do repair on the near sub, what value do I use now. I'm thinking the MVA they gave me minus 30%???? But I don't have a good idea for that value. Thats what I'm asking you Util guys. What do you use for your lowest fault MVA.

That is the reason I don't agree with the utility furnishing infinite source for arc flash studies. Too many assumptions to make!

It would not be unusual for the actual fault duty to be 40 to 50 % of infinite value and a weak source could be even less than that. I have seen them in the 25% range as an exception.

As for the lowest, I have always (to date) been able to get the actual numbers and work from there.

I have been working on something along these lines for a while.

I can pretty easily find the maximum point for arc flash using the Lee method and standard relay curves. I have yet to tackle the IEEE 1584 calculations and breaker curves with all of the variables involved.
The method that I am using involves putting the relay curve equations directly into the arc flash equation, then reducing this to per unit values, using system voltage and the breaker or relay long time pickup as references. The resulting equation can be graphed directly opposite the relay curves on a TCC.

Hopefully I will be able to get back to this soon.

Utility fault strategy

One strategy I've used and had luck with in obtaining actual fault current availability is to request the primary fault currents and X/R. If you've visited the facility to be studied and can determine the transformer properties you can model the service transformer and thus the secondary fault values. Many utilities don't track the impedance of the service transformers and the secondary fault values. But most can provide the fault values for the primary system.