This week’s question was submitted by one of our regular contributors. It is a good one but will take a bit of thought as the introduction is quite detailed.

Although maybe not a normal condition, some equipment can have multiple sources of short circuit current under special conditions. Two examples include:

Automatic Transfer Switch
Automatic Transfer Switches (ATSs) (when both utility and generator power are present). If an arc flash occurs when both sources are present, the fault current for the arc flash will be higher than with only the normal or generator sources.

It is possible that both sources can be present during commissioning, testing, or when performing the ATS/Generator exercise test. The ATS/Generator exercise test simulates the loss of utility power, starts up the generator, and transfers the load from utility to generator.

Double Ended Substation
Another example is a bus tie and main circuit breakers for double-ended substations. For the main and ties of a double ended substation, there are times that there can be sources of fault current available on both the line and load side of main and tie circuit breakers.

This week’s question:
When you perform arc flash studies, do you calculate the arc flash energy at the Main and Bus Tie breakers for double ended substations and at ATS locations where BOTH SOURCES could be available (although technically not paralleled) for these special operating conditions?

Due to character limits with the forum question posting, the above question was shortened to:

Do you include a case with both sources of fault current at an ATS and Doubled Ended Sub Mains and Ties?

Yes
No
It Depends (please elaborate)
Not my job/no opinion

I voted no because I have not been doing that. I'm not sure how to model that in an open transition ATS. I suppose you'd have to change it to a closed transition or add a bus that ties the normal and emergency together? I don't think that would provide accurate results as it is not exactly the same as what you're describing.

We have many double ended Main-Tie-Main boards on campus too but the mains and tie are all interlocked in some form or fashion (electrically or via Kirk-keyed locks) so as not to allow closing the two sources together through the tie. You'd have to ignore that and close the tie to model what you're describing. I've seen a few other engineer's studies with MTM's and non of them have created that scenario (MTM breakers all closed together) due to the interlocks.

EDIT: I also have study performed by one of the "major" gear manufacturers with both a double ended switchboard and with multiple ATS's on the project and their model does not account for the scenarios you're describing either.

The more I think about this, the less I think I understand the question.

Assuming the ATS's are open transition and the MTM is interlocked then for all practical purposes it is physically impossible for there to be a bolted 3 phase fault. So I must assume you are only talking about an arcing fault from between the E and either the load (when connected to normal) or normal in the ATS? Is that what you're asking?

What about in the MTM setup? Same deal. Bolted 3 phase fault is physically impossible so arcing fault between line and load of the tie breaker or those buses?

Am I overthinking this and you're just asking if we consider scenarios of load on normal and load on emergency and what is the worst case? That seems too simple and everyone should be doing that as standard modeling procedure. That is what the scenario tools or functions are for in modeling software.

It is a complex one. The way I understood it, if an arc flash occurs at lets say an ATS for example, whether it is open or closed transition the issue is there may be a case where there are two live sources available in the ATS even though the load is only connected to one source. The arc flash could escalate and involve both sources.

Although closed transition is another good case.

I will re-confirm with the source of this question to see if this is correct - but I believe so.

I witnessed an arc flash on a double ended substation when both mains were tied in with the tie breaker closed. I don't believe that scenario had been modeled in the arc flash study, but obviously it should have been as it actually happened.

There were supposed to be interlocks in the breakers that prevented that scenario, but during maintenance the wrong breakers, without the extra interlocks present, were put in service.

I will often model a closed tie breaker in a double ended M-T-M installation that is not intended for normal parallel operation, based on the operating methods of the facility.

We had one customer where the transfer from live Source A to live Source B was performed manually, there were no key interlocks. But, this was only done after all of the branch/feeder breakers had been opened so the item seeing two sources was the tie breaker compartment. Not modeling any motor contribution affected the incident energy almost as much as the paralleled sources.

Modeling an ATS with two live sources is quite common, however, I cannot remember ever modeling both sources contributing into a single fault.

If they are kirk-keyed or otherwise interlocked do you still do this? Someone would really have to be trying to close everything together by having extra keys made or locks removed. If you model this then I suspect your available fault currents and your incident energies are way high causing workers to adhere to unnecessary overkill practices, maybe. If that's OK then why even bother to do the models and just give everything "Danger" (I know, separate discussion) or "over 40 cal" or "Shutdown" or whatever labels and require all equipment be 100 or 200KAIC?



Yes, you just select what position the switch is in while keeping both sources in service. I don't think it affects any calculations though. At least not in SKM as it is not calculating both sources contributing to a single fault like you say. At least I'm pretty sure it doesn't. If I get time I'll run it both ways, with the generator in service and not in service and see if I get any different numbers at the ATS. I don't think I will as I don't think the software assumes the normal will fault into the emergency side.

My question is why do you need to perform live work inside an ATS that has utility and generator power present?

The IE from the generator and from the utility at an ATS will no doubt be very different. My guess would be the generator would have the higher IE because it has limited fault current and the time element in the IE calculation will take the dominant role compared to the higher utility fault current.

Closed transition transfers are typically under 100mSeconds. I would just ignore this as a possibility. If I had to work inside the ATS I would then label it whichever is greater, the utiltiy IE or the generator IE.

What do you then do with every panelboard downstream of the ATS, label it the higher of the two?

For ATS's and Panelboards supplied by ATS's, I post the utility IE, along with a cautionary note that "Exposed work while on generator is prohibited".

This way the label represents the real world day to day, and anything outside of that can wait until the utility comes back.

For the double ended substations with a bus tie, it does not matter if there are kirk-key interlocks. If both mains are closed, then there will be a source of fault current on the bus tie line and load side terminals. Double the fault current, possible double the arc flash energy. An example where this could be would be measuring voltage in the bus tie cubicle or racking in the bus tie breaker when both mains are closed.

Address the concern up front with the owner / end user, some educated owners have standards regarding this and it relates it to a couple of things mostly the equipment type, the transfer / parallel time [ ex. 100ms or less] and how the equipment will be operated under maintenance, etc…
Some owners recognize and understand the conversation and provide direction verbally, in an e-mail or in their owner standards that you do not need to evaluate the closed transition for arc flash only for equipment evaluation [short circuit] and to label for the normal operation condition only. These types of owners typically understand the hazard and when switching for maintenance will also typically do so with remotely activated ATO schemes not while in front [or behind] the gear.
Sometimes we are only requested to provide the study and associated spreadsheet with labels printed by others in that case provide all scenarios and allow them to label appropriately.
When the owner says, I need an arc flash study and that’s it starts after review of documents, or lack thereof, then a conversation to set the dials for the project on how equipment will be labeled, sample label, etc. so there are no questions at the end of the project. If they are still not sure, then the label typically reflects the closed transition state.
With transfer switches, and the emerging arc reduction technologies there is the other issue of how do you label or do you provide multiple labels?? Sometimes the hazard on the normal utility may be 1.2 – 8cal/cm^2 but on the generator, could exceed 40 due to the decrease of available fault current and the circuit breaker tripping characteristics. For one’s that want to label for the worst case this may always result in “unapproachable� equipment based on the results when truly 90% of the time it is not in that state, and some say providing multiple labels can create confusion. One owner I know labels only for normal condition and their work practice is to lock out the transfer switch [from transfer] if working live downstream, and if work is required while on emergency they refer to their SCCAF report that list the I.E. for alternate operating conditions [ex. Generator, Arc reduction technologies, etc.] for the associated equipment and associated downstream connected equipment...

Unfortunately, there is no easy answer for this one…

This is obvious. What is not obvious, to me anyway, is how to figure the incident energy. Do you merely add Source 1 to Source 2 or is it more complicated than that, assuming Source 1 and Source 2 are arcing between one another?

I think it is extremely unlikely you'd have a arcing fault on Source 1 and at the same time Source 2 on either side of the breaker but not involve each other in the event thereby rendering simply adding your results together and calling it a day inaccurate. Something really catastrophic is happening in the first place if Source 1 and Source 2 are both involved at the tie breaker.

These types of circuit to circuit faults are known as "cross country" faults. Doesn't your software model these faults along with the sequential clearing?

If the equipment allows continuous paralleling or closed transition transfer then I run a case for that scenario. If the equipment faults during the closed transition, it would see contribution from both sources.

Keep in mind some of the statements made previously defy the mathematics involved, and the analysis is not quite as complicated as one would think. With a fault in the tie assuming that somehow we get into a double bolted fault scenario, this would be identical to the short circuit current calculated with the tie closed. No need to figure out how to "add" incident energies. The software does this for you.

Second although recently a much improved method for doing this was done and presented at ESW, if the incident energy analysis indicates that the overcurrent protective device is in a definite time condition ("instantaneously" tripping or else nothing is tripping and we're just going to 2 seconds), then as short circuit current increases, incident energy increases as we would expect it to. However in the vast majority of cases for distribution-level substations such as almost all double ended substations, we are operating on an inverse time curve whether the protective device is a fuse or a circuit breaker. Thus as short circuit current increases, the arcing time decreases and in the vast majority of practical cases I have actually seen, even though arcing power is increasing, trip time is decreasing at a faster rate and the net result is a net decrease in incident energy.

The net result of the action of inverse time curves means that in the "double 3-phase arcing fault" case, incident energy will be less at the tie instead of greater. Similarly even though short circuit (and arcing) current is higher with both mains and the tie closed, the incident energy itself will be less. In fact the worst case is always with one main and one tie closed since at that point we have the greatest amount of fault current contribution from motors and generators with the highest impedance.

In any case, these operating cases are not the norm for most plants. It is best in my opinion to calculate incident energy considering the normal operating condition which covers incidents 99% of the time and to use that for label purposes. The other potential cases should be analyzed and part of the report to show what could happen under various scenarios so that there is a convenient source for looking up what happens with the various alternative operating (but not typical) operating conditions. There are arguments made that during "emergency" conditions personnel may not bother looking anything up and this might be true but again, this is a very marginal situation in the first place.

It should be recognized that "overprotecting" is not the same as being "conservative". If "overprotecting" causes PPE to increase enough to cross thresholds such as the 12 cal/cm2 threshold, it triggers additional safety hazards such as wearing thicker gloves, using much darker face shields, and increasing risks of heat-related illness as we go from single layer to multi-layer PPE. And at some point when the incident energy gets so high that either PPE is simply not available, then we enter a condition where all kinds of crazy work rules and approaches often happen, especially when plants refuse to perform necessary maintenance on equipment out of fear, consequently again increasing risks unnecessarily. These all create greater safety risks, which is directly counter to the principle of being "conservative" in terms of professional engineering. Thus labelling for the absolute worst case scenario when doing multiple scenario analysis may actually be increasing risks to plant personnel by pushing them into greater PPE.

Consider this scenario: ATS is open with utility sourcing one side and a not yet in sync generator sourcing the other. The fault occurs when the two sources are 180 degrees out of phase, and grows to involve contributions from both sides. Is this really the same as a fault with the tie closed and the sources in sync?

If they are synchronized them Isum=Ia+Ib for any phase by superposition. If they are 180 degrees out of synchronization then obviously we will have Isum=Ia-Ib. For any other case we have to calculate the in phase and out of phase values and sum/difference appropriately. So the synchronized case is actually the worst case in terms of available fault current.

With an arcing fault it extinguishes and then reignites at each zero crossing. Having two currents out of synchronization from each other could potentially cause up to 4 points where current passes through zero, lessening the arcing time and consequently the arc flash relative to the synchronous case.

Just wanted to respond and say thank you for taking the time to respond. I found this looking for a parallel feed of a bus and generator testing scenario. Thanks!