Multiple sources feeding a bus

I am an engineer doing an arc flash study using IEEE-1584 calculator (manual/excel based). I have come across a problem. Looking for assistance/clarification on my doubt.

Let us assume, there are multiple sources feeding a bus (Generator, Motor and Utility) hence also will contribute differently to the bus bolted fault level. On the occurence of an fault at this bus the individual breakers would trip the 3 sources independently depending upon the fault current seen by individual relays and the consequent breaker trip+operating time (depending upon each relay setting). For such a case how one should calculate the net incident energy at the operating personnel body in order to calculate the hazard category?

Not sure if I follow your layout exactly, but, the way I do this is

1) Assume maximum fault current from all sources that would be contributing, then look up clearing time. Calculate IE.

2) Assume lowest possible fault current from whichever source has the lowest fault current, assume no motors running at time of fault, then look up clearing time. Calculate IE.

Use whichever value of IE is highest.

Generators are often a particular problem has they have high impedence and low fault current values. Big problem with molded case, less of a problem with electronic trip.

You have to calculate each time period separately, then add the incident energies. Each time period being each length of time with a different arc current. The first would be with all sources. The next would be after the first source OCPD tripped.

You can't treat each source separately because the arc energy from different currents are not additive. The arc energy from a 2 kA arc is more that twice the energy from a 1 kA arc.

To jghrist,
Is there a time period till which I should calculate the accumulated incident energy. I have read in some literature that you should calculate till the 'cleared fault threshold limit', which is the portion of Total Arcing Fault current at the Bus that needs to be interrupted by protective device to extinguish the arc. Default value is sometimes taken to be 80% of Total Arcing Fault Current.

And just to verify based on your reply, The Hazard Category will then be defined based on 'Total Accumulated Incident Energy'?. Wont the results via this be conservative?

Haze I think my question was not very clear to you. As rightly understood by jghrist, the incident energy and the arcing current will be different, as different branch OCPD will be tripping at different point of times.

If you consider the maximum fault current from all sources and trip time to be of the last OCPD, then you would definitely be overassessing your hazard category levels.

There may be a practical limit to the minimum current that would sustain an arc, but I don't believe that it is given in either NFPA 70E or IEEE-1584. Unless you can come up with an industry accepted source, I would consider that the arc stays until all sources are tripped. The 80% figure seems suspect. If you had a 50kA arc fed 40kA from a utility source and 20kA from a local generator, I wouldn't expect the arc to extinguish when the generator tripped off if the utility source was still intact.

Got your point.

I am doing manual calculations for calculating incident energy levels.
As with case of motors of considering contribution for about 5-6 cycles (on conservative side), is there an off-hand number of cycles that you should take into consideration for generators (may be based on its size)?

I haven't really investigated this, but my first thought is that a generator subject to arc flash could take a very long time to trip. If you expect arcing current to be 30 to 70% or fault current, and the OCPD is a molded case breaker, you could be in for a long wait.

As the fault begins, the generator will lose speed as it attempts to put out more current. The governor increases the fuel rack. The fuel rack is eventually full throttle and the engines ability to produce HP is at maximum. Arcing current is not likely to be 15 to 20 times of breaker rating, so operation will be in the short or even the long time segment of the trip curve.

IEEE, as well as other investigators, tell us that the arc will not sustain itself indefinately, maybe not even 20 cycles. When the arc extinguishes, the engine could still be running. With the arc out, the low impedance short may be gone, not necessarily still in one piece but not faulted any longer either. The engine and generator is back to full speed and full voltage.

Point is that the OCPD may never trip at all. If you have a large generator with digital trip settings, you can probably set the trip level low enough. But most of the less than 1MW generators a coming with molded case breakers.

I believe modern software designed to calculate arc flash exposure takes into effect the clearing time of each device based on the fault current available from each device, and sums the contributions. The trip times are automatically adjusted to include reduced current due to arc resistance, per IEEE 1584.

For calculating Arc Flash Exposure on distribution system to be cleared by equipment located downstream of where the sources converge, it would be prudent to calculate the exposure multiple times with each of the multiple sources turned off in any arrangement that is likely to occur. The reason for this is that device clearing times will vary based on available fault current and the highest calculated exposure may not coincide with the highest available fault current.

This generator question has me perplexed.

SKM software models generator decrement by allowing the generator output to be reduced to X% at Y cycles.

I arrive at the %decrement from the manufacurer's data

My problems with this approach:

1. Generator decrement is based on a bolted current, which may only occur very briefly.
2. The study scenarios involve 1 to 3 machines in parallel, which may have higher fault clearing capability than the sum of their parts.
3. These generators are self excited, so depending on the response to a high impedance fault, may not sustain an arc for anything close to 2 seconds due to loss of excitation.

Does anyone have any insight on how to accurately model generator(s) response to a high impedance fault?

I hesitate to call these abnormal conditions or scenarios that needn't be studied. A critical type facility seems just as likely to require live work
running on generator power as when on the utility.

Thanks

Dual infeed to fault problem

I have come across this problem a few times now and it seems most significant when there is more than one infeed sharing contribution 50/50 with very different clearance times.

For example, i have a board with total fault of 19kA with two main infeeds, one at 9.6kA tripping in 1.75s and the other at 7.07kA tripping in 6.421s.

If i base the incident energy calcs on 19ka for 6.421sec i get 864.2J/cm2.

However, my approach has been to do the calulation in two stages:-

Stage one - total fault current 19kA, clearance time 1.075sec giving 144.35J/cm2

Stage two - interconnector has tripped, total fault current is now 9.4kA and clearance time is 5.346sec (that is 6.421sec minus 1.075sec as time has already elapsed there only remains 5.346sec before the other breaker will clear the fault) this gives 299.1 J/cm2.

Having considered the two stages, the full level of energy is the sum 144.35 + 299 = 443J/cm2? This is clearly a lot less than the original value of 864J/cm2.

There is a PDF atached which should illustrate this, note that the values are to highlight the problem only, they are artifically high!

I would be very interested to hear anybody's comments on this approach. I believe it reasonably captures the dynamic events and changes in energy over the clearance time period.

Phil

Philip-

Performing a 2 stage calculation is the correct way to do this. The incident energy is not linear, but the current values are. I came up with similar numbers, but first, a few questions:

Do your currents represent bolted currents or arcing currents? The difference is minimal at this voltage, but trip times should be based on arcing currents.

I see motors on the one line, their contribution should also be included, just as your other parallel sources are. This will result in additional stages to your calcs, depending on the duration of their contributions.

Is there any reason that you didn't terminate your calculation at 2 seconds? There are cases when a worker may be exposed to the energy for longer, but the concensus seems to be that a worker can escape or is blown clear within this interval. There are a few threads on this subject in the forum.

Finally, with multiple sources, it may not be necessary to sum all of the fault sources for the duration of the event. After the majority of the contribution has been removed from the bus, (use 80%) there is a point where the arcing stops, even though there is still a contribution present.

Good luck!